On 3 September 2026, the European Commission’s guidelines on the EU Forced Labour Regulation were published in the Official Journal as C/2026/4637. They were adopted on 26 June 2026 and presented on 30 June alongside the launch of the Forced Labour Single Portal. Together with an implementing regulation that went into force in May, they are the clearest picture yet of how this law will actually be enforced.

That picture is worth a close look, because the EU Forced Labour Regulation behaves differently from almost every other supply chain rule procurement teams have absorbed in the past three years. There is no periodic report to file. No declaration, no certificate, no registry entry, no annual statement. Nothing falls due on a recurring date.

Instead, there is one question, and it arrives without warning. An authority writes to you about a specific product and asks what you have done about forced labour risk in that product’s supply chain. You have 30 working days to answer.

The good news is that 30 working days is a generous window if the answer already exists, and an impossible one if it has to be assembled from scratch. The regulation applies from 14 December 2027, which leaves roughly fifteen months to make sure you are in the first category. This article is about what that actually requires, and why it is a different build from the one most compliance programmes are running.

What the regulation actually does

Regulation (EU) 2024/3015 entered into force on 13 December 2024 and applies from 14 December 2027. Article 3 is short:

“Economic operators shall not place or make available on the Union market products that are made with forced labour, nor shall they export such products.”

A few features of that sentence matter more than they first appear.

There is no value threshold and no de minimis. There is no exemption for small and medium enterprises. It covers imports and exports alike. And it reaches forced labour used at any stage of extraction, harvest, production or manufacture, anywhere in the supply chain, which means a single component can bring a finished product into scope.

This is product legislation, not reporting legislation. The obligation is an obligation of result: the product must be clean. Enforcement runs through national competent authorities, which Member States were required to designate by 14 December 2025, and through customs, which can suspend and then refuse release of affected goods at the border.

One qualification on the no-paperwork point is worth knowing. Article 27 allows the Commission to designate, by delegated act, products whose economic operators must provide identifying information to customs: a product description and identifiers, plus the name and contact details of the manufacturer and of the suppliers. That is a data duty attached to a shipment rather than a report on a calendar, and it only applies to products the Commission specifically lists. Until products are designated under that power there is nothing to supply, which makes it a watch item rather than a workstream. It is also one more reason to hold manufacturer and supplier identity as structured data rather than as attachments.

One further point deserves emphasis, because a great deal of commentary gets it backwards. The burden of proof sits with the authority, not with you. The Commission’s guidelines state that the lead competent authority “bears the burden of establishing a violation of the forced labour ban” and “must provide credible evidence” that forced labour was used. You are not required to prove a negative about your entire supply base. You are required to be able to respond credibly when asked about a particular product.

It is also worth knowing what a decision does not do. Withdrawal reaches products still being made available on the market. It does not reach products that have already been sold to end users, so this is not a consumer recall regime.

The artefact to design around: an Article 17 request

Most preparation advice for this regulation stays at the level of policy. The more useful exercise is to look at the actual piece of paper that will land on someone’s desk, and work backwards from it.

In most cases, before any formal investigation, there is a preliminary phase under Article 17. The lead competent authority requests information from the economic operators under assessment about the actions they have taken to identify, prevent, mitigate, bring to an end or remediate risks of forced labour in their operations and supply chains. You have 30 working days to respond. The authority can skip that request where making it would jeopardise the outcome of its assessment, so this is the usual route rather than a guaranteed one. The authority then has 30 working days from receiving your information to decide whether a “substantiated concern” exists, which the guidelines define as “a reasonable indication based on objective, factual and verifiable information” to suspect that a product was likely made with forced labour.

If it concludes there is a substantiated concern, the matter moves to a formal investigation under Article 18, where the timescales change again: there the authority sets a deadline of at least 30 and no more than 60 working days for submissions, and must take into account the size and resources of the operator, including whether it is an SME, when considering extensions.

Those two stages are frequently conflated in summaries of this law. The distinction matters in practice, because the preliminary phase is the one where a good answer can end the matter before it becomes an investigation. That is the stage worth designing for.

Why this is a product question, not a supplier question

Here is the part that catches well-run programmes out.

Almost all supplier compliance data is supplier-shaped. A vendor register. A signed code of conduct per supplier. An ESG questionnaire per supplier. A risk score per supplier. If someone asks “what do we know about this vendor”, the answer is one row and a folder of attachments.

An Article 17 request arrives attached to a named product. The question is framed around your operations and supply chains, but the thing under assessment is a product, and a credible answer has to trace that product’s chain: from the finished item to the suppliers who contributed to it, to the specific sites those suppliers used, to the inputs that came from somewhere upstream. A supplier register cannot do that walk. It knows who you buy from. It does not know what is in what.

This is the single largest readiness gap we see, and it is a data modelling problem rather than a policy problem. The fix is to make the link between supplier, site and product an actual field in your supplier information management records rather than institutional knowledge held by a category manager. Teams that have already invested in procurement data management and supplier master data governance have a considerable head start here, because the hard part is not collecting the data once. The hard part is keeping the mapping accurate as the supply base changes.

A readiness self-test

For any product line you sell into or out of the EU, can you produce the following within 30 working days, without a special project?

Question Where the answer should already live
Which suppliers contribute to this product, and which of their sites? Supplier master data, with a product and site link
Which inputs come from an area or sector with documented forced labour risk? Risk-based supplier segmentation and country or sector risk flags
What did we ask these suppliers, when, and what did they answer? Supplier portal and evaluation history
What did we do when an answer was weak or missing? Corrective and preventive action records
Can we show that trail without rebuilding it from inboxes? Reporting and audit trail
Who owns the response, and how quickly can they assemble it? A named owner and a cross-department process

A team that answers yes to all six is in good shape, whatever its policy documents say. A team that answers no to three or more will find 30 working days very short, which is the practical argument for starting now rather than in 2027.

If most of those answers currently live in spreadsheets and email threads, that is the thing to change first. The problem is not that spreadsheets hold the wrong data. It is that they cannot produce a defensible history of who was asked what and when, which is precisely what the preliminary phase turns on. This is the same case for replacing spreadsheets in supplier management that quality teams already know, arriving now from a compliance direction.

What due diligence buys you, and what it does not

Be clear on this, because the marketing around forced labour compliance is not.

The regulation does not impose a due diligence obligation. The Commission’s guidelines are explicit that it operates “whilst not imposing any specific due diligence obligations on companies”, while recognising due diligence as “a useful tool to address forced labour in supply chains”.

There is also no safe harbour. The guidelines state plainly that providing information “does not guarantee exculpation for the economic operator or that the product will not be subject to a ban decision”. No certificate, audit report or third-party attestation changes that. Any vendor promising forced labour compliance as a product is selling something the regulation does not offer.

So what is the return on doing the work?

Three things, all real. First, due diligence evidence is the material you put in front of the authority in the preliminary phase, and it is one of the inputs weighed when deciding whether a substantiated concern exists. Second, the guidelines note that where an operator demonstrates within the deadline that it is taking appropriate steps to address a forced labour situation, the authority may allow a reasonable period to complete them. That is discretionary rather than guaranteed, but it does not become available at all to a company that cannot show what it is doing. Third, and least discussed, the same evidence base serves several regimes at once.

That last point is the strongest commercial argument. The supplier data that answers an Article 17 request overlaps heavily with what you already collect for CSDDD and for ESG and CSRD requirements in onboarding. Collecting it once and reusing it is the difference between a manageable programme and four parallel questionnaires landing on the same supplier, which is the fastest way to depress response rates across all of them.

Four things worth doing in the next quarter

Deliberately capability-shaped rather than calendar-shaped, because the deadline is far enough away that a week-by-week plan would be theatre.

Start with products, not suppliers. Pick the two or three product families with the most exposure, judged on where inputs originate and which sectors they sit in. Map those properly, end to end. A complete map of three product families is worth more than a shallow questionnaire sent to every vendor, and it tells you how much work the full picture will be.

Make the supplier-to-product link a field. If the only place that knows which supplier feeds which product is a category manager’s memory or a spreadsheet tab, that is the first thing to fix. Everything else depends on it.

Ask once and reuse. Fold forced labour questions into the onboarding questionnaire and the existing evaluation cycle rather than launching a standalone campaign. Suppliers have a finite amount of attention, and evaluation fatigue is the reason well-designed programmes stall. Where you operate across borders, multi-language supplier onboarding matters more here than usual, since the suppliers closest to the risk are often the least likely to work comfortably in English.

Let the trail build itself. The audit trail should be a by-product of running the process, not a document someone writes afterwards. If a colleague has to reconstruct the history from inboxes, you do not have 30 working days, you have 30 working days minus a forensic exercise. Automated follow-up and continuous monitoring rather than annual reviews both help, because they leave a dated record as they go.

For regulated sectors the case is sharper still. Teams running supplier relationship management for medical devices or for pharma already maintain product-level supplier traceability for other reasons, and extending it is a smaller step than building it.

What exists today, and what does not

Worth being precise, because plans have already been built on things that are not there yet.

Available now. The guidelines, adopted 26 June 2026 and published in the Official Journal on 3 September 2026. They are a Commission Notice and non-binding, but they carry weight, not least because the Commission itself acts as lead competent authority for cases outside the EU. The Forced Labour Single Portal has been live since 30 June 2026. Commission Implementing Regulation (EU) 2026/903 of 24 April 2026, in force since 17 May 2026, specifies the details and functionalities of the system competent authorities and customs will use to exchange information, which is a reasonable signal of how coordinated enforcement is intended to be.

Not available yet. The database of forced labour risk areas and products under Article 8. The Commission’s own page for it still carries the label “Database in development”. When it does arrive it will be indicative and non-exhaustive by design, so it will inform your risk assessment rather than settle it. Any guidance telling you to check your products against the database today is describing something that does not exist, and any plan that waits for it is fifteen months of waiting.

Coming up. The Commission is running a webinar series through autumn 2026: solar on 13 October, textiles on 20 October, electronics and semiconductors on 12 November, agri-food on 17 November, automotive on 19 November, and fisheries on 24 November. If one covers your category, it is a direct read on how the Commission thinks about risk in your sector, which is useful input for deciding where to start.

The encouraging part

Strip away the novelty and this regulation asks for something procurement teams have been building towards anyway: knowing what is in your products, knowing who made each part, knowing what you asked and what came back, and being able to show it quickly.

Teams already running structured supplier evaluation with a working supplier risk management process have most of the machinery. What usually needs adding is the product dimension and the speed of retrieval. Both are ordinary data problems with ordinary solutions, and fifteen months is a comfortable runway for ordinary data problems.

The teams that will struggle in December 2027 are not the ones with gaps in their supply chains. Every supply chain has gaps. They are the ones who cannot describe their own supply chain quickly enough to have a useful conversation about it.

If you want to see what product-level supplier traceability looks like in practice, a supplier performance pilot on a single product family is a low-commitment way to find out, or you can book a demo and we will walk through it with your own categories.

Frequently asked questions

When does the EU Forced Labour Regulation apply?
It applies from 14 December 2027. The regulation entered into force on 13 December 2024, and Member States were required to designate their competent authorities by 14 December 2025.

Does the EU Forced Labour Regulation require due diligence?
No. The Commission’s guidelines confirm it does not impose specific due diligence obligations. Due diligence is still practically important, because the evidence of it is what you submit during the Article 17 preliminary phase, and one of the inputs the authority weighs when deciding whether a substantiated concern exists.

Is there a safe harbour if we can show strong due diligence?
No. The guidelines state that providing information does not guarantee exculpation or that a product will avoid a ban decision. Certificates and audit reports are not a safe harbour either. What good evidence buys you is a faster, more credible response and, at the authority’s discretion, potentially time to complete remediation steps.

How long do we have to respond to an information request?
30 working days in the preliminary phase under Article 17. The authority then has 30 working days from receiving your response to decide whether a substantiated concern exists. If the matter moves to a formal investigation under Article 18, the authority sets a separate deadline of at least 30 and no more than 60 working days.

Are small and medium enterprises exempt?
No. There is no SME exemption and no value threshold. SMEs do receive specific consideration: Article 10 provides for support measures for SMEs, and the Commission’s guidelines note that SMEs can seek help from a national contact point on how to engage with the lead competent authority. Authorities are directed to weigh size and economic resources when prioritising enforcement, and must take them into account when setting deadlines during an investigation under Article 18(4).

Can we check our products against the EU forced labour risk database?
Not yet. The Article 8 database is still in development as of late September 2026. When published it will be indicative and non-exhaustive, drawing on sources such as ILO reporting, so it will support your own risk assessment rather than replace it.

What happens to products found to be made with forced labour?
The authority prohibits placing or making them available on the EU market and prohibits export, orders withdrawal of products still being made available, and orders disposal. Products are disposed of in line with the waste hierarchy, by recycling where possible or otherwise rendering them inoperable; perishable products are donated for charitable or public interest purposes. Operators generally get at least 30 working days to comply for non-perishable goods and 10 working days for perishable goods. Products that have already reached end users are not withdrawn.


Sources

Related reading

Picture a supplier that misses its on-time-in-full target for the third month running. The quality engineer opens a corrective action. The category manager wants to know whether the contract allows a service credit. Nobody can find the qualification note from eighteen months ago that approved the supplier “subject to confirmation of second-shift capacity”. Every stage of the lifecycle did its job. The failure happened in the gaps between them.

That pattern is the reason for this article. Most supplier lifecycle management programmes are assessed by listing their stages and checking that each one has a process, an owner and a screen. The better test is to look at what moves from one stage to the next, and what gets dropped on the way.

SLM, SRM and SPM are three different scopes

The three terms get used as synonyms. They describe different jobs, and mixing them up is one reason handovers go unowned.

  • Supplier performance management (SPM) measures. It defines KPIs, collects data, builds supplier scorecards and flags deviations.
  • Supplier relationship management (SRM) manages the relationship itself. CIPS defines it as “the management and maintenance of the relationship between a buyer and supplier” and notes that the type of relationship depends on the criticality of what is being bought. In practice SRM is selective: strategic suppliers get QBRs and joint plans, the long tail does not.
  • Supplier lifecycle management (SLM) covers the whole span, from the first internal need to the last day of the relationship, for every supplier you buy from. Its real subject is continuity: the same supplier record, the same obligations and the same accountability carried from sourcing through to renewal or exit.

SPM and SRM live inside SLM. If you want the finer distinction between the first two, we covered it in the difference between SPM and SRM. For a stage-by-stage walkthrough of the lifecycle itself, see full supplier lifecycle management from sourcing to offboarding. This piece picks up where those leave off: the joins.

Why modules are the wrong unit of evaluation

Every supplier management vendor can show you an onboarding screen, a scorecard and a CAPA form. A module checklist will almost always come back green. It tells you very little about whether the qualification risk, the contract KPI and the corrective action are the same thread or three unrelated records.

The standards that supplier quality teams work to are written as chains, not as separate boxes. ISO 9001:2015, still the basis of most certificates while the transition to ISO 9001:2026 runs, shows this in clause 8.4.1, which puts the whole sequence into one requirement: the organisation shall “determine and apply criteria for the evaluation, selection, monitoring of performance, and re-evaluation of external providers”, and retain documented information of those activities and any actions arising from them. Clause 8.4.3 goes further and requires you to communicate to the supplier how its performance will be controlled and monitored. In other words, the performance regime is supposed to be known at the point of contract, not invented afterwards.

IATF 16949 makes the dependency explicit for automotive. Clause 8.4.2.5 asks organisations to determine the priority, type and timing of supplier development actions based on defined inputs: performance issues from supplier monitoring, second-party audit findings, third-party QMS certification status and risk analysis. Supplier development is defined by what it receives from earlier stages. If those inputs do not arrive, the clause cannot be met, however good your development process looks on paper.

The commercial side sees the same thing. World Commerce & Contracting put it plainly in June 2026: “Value leaks between handoffs.” In an earlier piece, WorldCC’s Tim Cummins cited the association’s own research identifying 11% value erosion in procurement contracts and described how procurement focuses on award, legal on risk allocation, and operations “inherit the consequences”. Each function is competent. The value disappears between them.

The lifecycle as a chain of handovers

The table below treats each stage as a supplier of information to the next one. For each stage it names the question the stage exists to answer, the data object it must pass on, and the failure we see most often at that join.

Supplier lifecycle stages, the data object each hands on, and the typical handover failure
Stage Question it must answer Data object it hands on Typical failure at the handover
Intake and needs capture What exactly do we need, and how critical is it? Requirement record with criticality, specification and budget owner Criticality is never recorded, so every later stage treats a sole-source part like office supplies
Sourcing (RFI, RFQ, RFP) Who can meet the requirement, at what cost and risk? Evaluated bids with weighted criteria and the supplier’s own commitments Commitments made in the bid (lead times, capacity, certifications) stay in the RFx file and never reach the contract
Qualification Is this supplier and site capable and compliant? Approval status per supplier and site, with conditions and open risks “Approved with conditions” becomes “approved”, and the conditions disappear
Award and contract What have both sides agreed to deliver and measure? Obligations: SLAs and KPIs with definitions, targets, remedies and review cadence KPIs are written as prose in a PDF, with no definition a scorecard could use
Onboarding Is the supplier set up to perform from day one? Complete supplier master record, contacts, documents, and the KPI set switched on Onboarding closes when the vendor number is created, before performance measurement is configured
Performance and risk monitoring Is the supplier delivering what was agreed, and is its risk changing? Scorecard history, risk signals and threshold breaches Scores are reviewed but a breach has no owner and no route into corrective action
Development and CAPA What will change, who owns it, and did it work? CAPA with root cause, actions, owners, due dates and verified effectiveness CAPA is closed on paperwork, and nobody checks the following months of scorecard data
Renewal or exit Should we continue, change terms, or leave? Full relationship history feeding the next sourcing decision, or a retained record at exit Renewal is decided on price alone and three years of performance and CAPA history go unread

Read the right-hand column again. None of those failures is caused by a bad stage. Each one is caused by an object that was produced and then not received.

Four handovers that deserve your attention first

Qualification to contract: conditions must become clauses

Qualification often ends in a conditional approval. The site passed the audit but the process capability study on a critical characteristic was marginal, or the supplier’s certification expires in five months. Those conditions are the most valuable output of qualification, and they are the easiest to lose. The fix is mechanical: every condition should become either a contract clause, a monitored KPI, or a dated follow-up task. If it is none of these, it will not survive the award.

Contract to performance: one definition, not two

This is where most programmes quietly break. The contract says 97% OTIF. The scorecard also says OTIF. But the contract measures against the requested date and the scorecard against the confirmed date, or the contract counts PPM at goods receipt while quality counts it at the line. Both numbers are correct and they will never agree. Because ISO 9001 clause 8.4.3 already requires you to tell the supplier how its performance will be monitored, the contract is the natural place to fix the definition once. The scorecard should inherit it, not reinterpret it.

Performance to CAPA: every threshold needs a route

IATF 16949 clause 8.4.2.4 names minimum monitoring indicators: delivered product conformity, customer disruptions at the receiving plant, delivery schedule performance and premium freight occurrences. Measuring them is the easy part. The handover question is what happens when one crosses a limit. If a breach sits on a dashboard until the next quarterly review, the scorecard is only reporting the problem. A good handover opens a supplier corrective action with a named owner, links it to the metric and the clause it breaches, and closes it only when later data confirms the effect.

CAPA and performance to renewal: the history must be in the room

Renewal is where the lifecycle is supposed to loop back to sourcing. It frequently becomes a fresh negotiation, run by someone who was not involved in the previous term, looking at price and volume. The scorecard trend, the CAPA record and the risk changes over the contract period are the strongest negotiating material a buyer has. They are only in the room if the renewal stage receives them automatically. The same applies at exit: ISO 9001 asks you to retain documented information of evaluations and resulting actions, and a clean audit trail is what lets you answer a question about a supplier you stopped using two years ago.

A handover audit you can run in an afternoon

You do not need a transformation programme to find out where your lifecycle leaks. Pick three suppliers: one onboarded in the last six months, one mid-contract, and one approaching renewal. For each of the seven handovers in the table, ask three questions.

  1. Does the object arrive, or is it re-keyed? If someone copies KPI targets from a contract PDF into a spreadsheet, the handover exists only as long as that person does.
  2. Is there a named owner on both sides? A handover with a sender and no receiver is a filing exercise.
  3. Can you trace it backwards? Start from an open CAPA and try to reach the contract clause, the qualification record and the original sourcing commitment. Count the systems and the people you had to ask.

Score each handover 0 (missing), 1 (manual) or 2 (connected and traceable). This is a diagnostic, not a benchmark, so the total matters less than the pattern. Most teams find one or two joins scoring zero, and those are almost always the ones where different functions own the stages on either side.

Better questions to ask a supplier lifecycle management vendor

Swap the module checklist for handover demonstrations. Ask the vendor to show you, live, with their own demo data:

  • A qualification condition that reappears as a tracked KPI or task after award.
  • A contract SLA and the scorecard metric that inherits its definition and target.
  • A threshold breach that opens a CAPA linked to both the metric and the clause.
  • A CAPA closed only after subsequent performance data confirmed the fix.
  • A renewal view that shows the full performance, risk and CAPA history for the term.
  • A backwards trace from any CAPA to the original sourcing event.

If the vendor offers AI across these steps, ask what it reads, what it proposes, and who approves before anything reaches a supplier. Our guide on how to compare AI features in SRM software sets out those questions in more detail.

How EvaluationsHub, InitiativesHub and Eva AI close the loop

We built our two products around the handovers, because that is where supplier data and accountability most often go missing.

InitiativesHub covers the execution side of the lifecycle: intake and needs capture, RFx (RFI, RFQ, RFP), supplier qualification, awarding and negotiation, and contract management with SLAs and KPIs mapped to the clauses they come from and deviation tracking. It ends with supplier onboarding and a handover to EvaluationsHub, so the obligations agreed in the contract arrive as the starting configuration for performance measurement rather than as a PDF.

EvaluationsHub takes over from there: weighted performance scorecards, supplier risk, ESG and CSRD data, CAPA, QBR management, a two-way supplier portal, and reporting with a full audit trail. Together the two cover the full supplier lifecycle, with renewal decisions drawing on the history EvaluationsHub has kept.

Eva AI sits on top of both and is designed to watch the contract-to-performance join in particular. It reads contract KPI obligations, compares them against live scorecard data, and proposes a CAPA with its reasoning when a deviation crosses a threshold you have defined. It is also designed to monitor risk signals against your acceptance frame and to recommend compliance workflows linked to the relevant clause. Every proposed action is held for human approval, with one-click approve, edit or dismiss, nothing goes to a supplier without sign-off, and each AI decision is logged in an immutable, exportable audit log.

If you want to see a handover trace on your own supplier scenario, book a demo.

Frequently asked questions

What is supplier lifecycle management?

Supplier lifecycle management (SLM) is the end-to-end management of every supplier from the first internal need, through sourcing, qualification, contract, onboarding, performance monitoring and development, to renewal or exit. Its purpose is to keep one continuous supplier record, with obligations and accountability carried from stage to stage.

What is the difference between SLM, SRM and SPM?

SPM measures supplier performance through KPIs and scorecards. SRM manages the relationship, usually with extra depth for strategic suppliers. SLM is the widest scope: it covers every stage and every supplier, and it contains both SPM and SRM.

Where do supplier lifecycle programmes usually fail?

Most failures happen between stages rather than inside them. The most common are qualification conditions that never reach the contract, contract KPIs defined differently on the scorecard, threshold breaches with no route into corrective action, and renewals decided without the performance history.

Does ISO 9001 require supplier lifecycle management?

ISO 9001:2015 does not use the term. Clause 8.4 does require criteria for the evaluation, selection, monitoring of performance and re-evaluation of external providers, retained documented information of those activities and resulting actions, and communication to suppliers of how their performance will be monitored. A connected lifecycle is a practical way to meet those requirements consistently.

By Prof. Dr. Bert Paesbrugghe, founder of EvaluationsHub and associate professor at IÉSEG School of Management.

Ask ten supplier quality teams how often they re-evaluate suppliers and most will say once a year. Ask why once a year, and the answers get thinner: it is what the last system did, it fits the budget cycle, it is what the auditor saw last time.

Annual is a habit rather than a rule. ISO 9001 asks for criteria to evaluate, select, monitor and re-evaluate external providers, and for controls proportionate to the impact of what those providers deliver. It does not name an interval. That silence is an opportunity, because the teams getting the most from their supply base are not evaluating everyone more often. They are evaluating far fewer suppliers on the calendar and far more of them on evidence.

This article sets out a cadence model you can defend to an auditor, explain to a supplier and actually sustain.

The mistake is having one clock

Most programmes run a single annual cycle that bundles three genuinely different activities together. Separating them is the single highest-return change available, because only one of the three costs your suppliers anything.

Clock one: signal monitoring. This runs on data you already own. On-time delivery, quality escapes, open nonconformities, invoice disputes, responsiveness to requests. No supplier has to fill in anything for you to see a trend forming. Because it is free at the supplier’s end, this clock should run continuously rather than annually. Teams making the shift from annual reviews to continuous supplier monitoring usually find the signals were sitting in their own systems the whole time.

Clock two: supplier-effort data collection. Questionnaires, certificates, self-assessments, document refreshes, audits. Every cycle spends a finite amount of supplier goodwill, and it is the only clock the supplier really feels. This is the one to ration, and the rest of this article is mostly about how.

Clock three: the relationship review. The conversation about direction, capacity, roadmap and joint improvement. This is a business rhythm rather than a compliance one, and it belongs with the suppliers where the upside justifies the time. A well-run quarterly business review drives supplier change in a way an annual form never will.

Once the clocks are separate, the question stops being “how often do we evaluate everyone” and becomes three sharper questions. What do we watch constantly? What do we ask for, and from whom? Who do we sit down with?

What regulators and accreditors already worked out

Setting inspection frequency under resource constraints is a solved problem in several industries, and the solutions converge on the same principle: frequency should be earned, not fixed.

Aerospace: intervals as a reward. In the Nadcap programme, run by the Performance Review Institute, initial accreditation runs for 12 months before a reaccreditation audit is expected. Suppliers with a strong record can then qualify for merit status, under which accreditations last 18 months, and later 24. The criteria are specific. For 18-month merit a supplier needs to be at least at its second reaccreditation in that commodity, carry no non-sustaining corrective actions from the current or previous audit, and stay within defined limits on delinquency days and nonconformance counts. The 24-month interval asks for more: two consecutive accreditations of at least 18 months each, no major nonconformances, and no more than seven cumulative days of delinquency.

The design is worth pausing on. A longer interval is not an administrative convenience granted to save the auditor time. It is a status the supplier earns, can see the criteria for, and can lose.

Food safety in the EU: frequency proportionate to risk. Regulation (EU) 2017/625 requires official controls to be performed regularly, on a risk basis and at appropriate frequency, with authorities adjusting the control effort to the risk and the level of compliance expected. The regulation is explicit that a good compliance record should mean lighter treatment, noting that operators with a good record of compliance should bear lower overall charges than non-compliant ones, as they should be subject to less frequent official controls.

Food safety in the US: risk-based frequency written into law. The FDA Food Safety Modernization Act established a mandated inspection frequency based on risk, requiring that all high-risk domestic facilities be inspected within five years of enactment and no less than every three years thereafter.

Three different regimes, three different industries, one shared conclusion. Nobody with real resource constraints inspects everything at the same interval, and nobody treats a good track record as irrelevant to how often you come back.

Your supplier programme has exactly the same constraints. It can use exactly the same logic.

A merit-based cadence model

Here is the model in one table. Treat the intervals as starting points to calibrate against your own category risk, not as universal truths.

Tier Who belongs here Clock two interval How a supplier moves up How a supplier moves down
Watch Critical or single-source suppliers, new suppliers in their first year, suppliers with an open major nonconformance 3 to 6 months Not applicable, this is the entry and recovery tier Not applicable
Standard Suppliers with real impact on conformity and a normal record 12 months Complete two consecutive clean cycles with no major findings and no overdue CAPAs Any major nonconformance, or two consecutive missed response deadlines
Merit Proven suppliers with a sustained record 18 to 24 months Two consecutive Standard cycles clean, all CAPAs closed on time, scorecard above threshold Same triggers as above, returning to Standard or Watch depending on severity
Light touch Low-impact, easily substituted suppliers 24 to 36 months, or on event only Not applicable Escalates if spend, criticality or risk exposure changes

Two design notes matter more than the numbers.

First, the tier follows impact on conformity, not spend. A low-spend supplier of a safety-critical component belongs above a high-spend supplier of office consumables. If you have not segmented yet, a risk-based supplier segmentation or a Kraljic-based split gets you most of the way in an afternoon.

Second, publish the criteria. A supplier who knows that two clean cycles earns a longer interval has a reason to respond well to the current one. A supplier who thinks the schedule is arbitrary has no such reason. This is the quiet advantage of the merit approach: it converts your evaluation cadence from a cost you impose into an incentive you offer, using nothing but information you already hold. Make the thresholds visible in your supplier scorecards, and if you are still designing the scoring model, our guide to building a weighted supplier scorecard covers how to set thresholds that hold up across categories.

Events should outrank the calendar

The strongest argument against pure calendar scheduling is that risk does not arrive in annual instalments. A re-evaluation triggered by something that actually happened is worth several triggered by a date.

Build a short, agreed list of events that pull a supplier forward regardless of when they are next due:

  • A major nonconformance or a repeat of a previously closed one
  • A change of ownership, site, or key sub-tier supplier
  • A sustained fall in on-time delivery or responsiveness on clock one
  • Entry into a new regulatory scope, a new market, or a new critical part
  • A disruption affecting their region, sector or logistics route
  • Loss or lapse of a certification you rely on

This is where clock one earns its place. Continuous signals are what make event triggers possible, and automated follow-up for supplier evaluation and risk turns a trigger into a request without anyone remembering to send it. Business continuity deserves a specific mention here, and it is worth telling your suppliers that continuity is a priority rather than discovering their assumptions during an incident.

Anything that triggers a re-evaluation should also close properly. A finding that leads nowhere teaches the supplier that findings lead nowhere, so each one needs a tracked supplier CAPA with an owner, a date and verification of effectiveness.

Spending the attention budget well

Think of supplier attention as a budget rather than a free resource. Every questionnaire draws on the same account, and the balance is shared across you, your colleagues in other departments and your supplier’s other customers. Cadence decisions are really allocation decisions.

Three moves free up budget without reducing coverage.

Collect once, use many times. Overlapping requests from quality, procurement and sustainability are the most common source of avoidable supplier effort. A single structured intake, through a supplier portal or a shared onboarding process, lets one submission serve several internal purposes.

Ask only for what changes. Much of a typical annual questionnaire repeats information that has not moved since last time. Pre-filling known answers and asking suppliers to confirm or correct turns a long form into a short one, which is a large part of avoiding evaluation fatigue in procurement teams and at the supplier’s end too.

Give something back each cycle. The most reliable way to keep response rates healthy over a long cadence is to return the supplier’s own scorecard with peer context and a clear view of where they stand. It costs you nothing you do not already have, and it changes the request from an obligation into an exchange. Consistency helps here, which is why reducing bias in supplier performance reviews matters as much for credibility with suppliers as for internal fairness.

Where an on-site visit would be the natural next step but the travel is hard to justify, digital supplier audits let you keep a meaningful interval without the cost that usually forces teams to stretch it.

Making the change without an argument

Changing cadence tends to raise two internal objections, and both have straightforward answers.

“Reducing frequency looks like reducing control.” It is the opposite if you say it properly. You are moving from a fixed interval applied uniformly to a risk-proportionate interval with documented criteria, continuous monitoring underneath it and event triggers on top. That is a stronger control story, and it is the same logic regulators use. Record the rationale, the tier definitions and the trigger list as documented information, and your reporting and audit trail shows an auditor a considered system rather than a habit.

“We cannot track different intervals for different suppliers.” This is a real constraint in a spreadsheet, where a single shared deadline is the only manageable option. It stops being a constraint once intervals, triggers and reminders are held in a system, which is a common reason teams replace spreadsheets for supplier management and the thing that makes differentiated cadence practical when you scale a programme past a few hundred suppliers.

In regulated sectors the tier boundaries sit tighter and the documentation bar is higher, though the structure holds. Teams working under device or pharmaceutical oversight will recognise it from supplier relationship management for medical devices and for pharma, and automotive teams from the customer-specific requirements behind a scorecard built for IATF 16949.

A sensible first step is to leave the cadence alone and split the clocks. Put continuous signals in place, agree the event trigger list, and let the evidence from the first two quarters tell you which suppliers have earned a longer interval. By the time you change the schedule, you will have the record to justify it.

If you want to see how tiered intervals, triggers and automated follow-up work in one place, you can start a supplier performance pilot or book a demo.

Frequently asked questions

How often should you evaluate suppliers?
There is no single correct interval. A practical model is 3 to 6 months for critical, new or recovering suppliers, 12 months as a standard cycle, 18 to 24 months for proven suppliers who have earned a longer interval through consecutive clean cycles, and 24 to 36 months or event-only for low-impact suppliers. Continuous monitoring of data you already hold should run underneath all tiers.

Does ISO 9001 require annual supplier evaluation?
No. ISO 9001 requires defined criteria for evaluating, selecting, monitoring and re-evaluating external providers, with controls proportionate to the impact of what they supply, and documented evidence. It does not specify a frequency. Annual review is a widespread convention rather than a requirement, so a documented risk-based cadence is fully compatible with the standard.

Is it risky to reduce supplier evaluation frequency?
It depends entirely on what replaces the removed cycles. Reducing frequency on its own reduces oversight. Reducing frequency while adding continuous signal monitoring and event-based triggers usually increases it, because you learn about problems when they happen rather than at the next scheduled date. This is the same reasoning behind risk-based inspection frequency in food safety regulation and merit-based audit intervals in aerospace accreditation.

What should trigger an off-cycle supplier re-evaluation?
Common triggers are a major or repeat nonconformance, a change of ownership, site or critical sub-tier supplier, a sustained decline in delivery or responsiveness, entry into a new regulatory scope or critical part, a disruption affecting the supplier’s region or sector, and the lapse of a certification you rely on.

How do you decide which suppliers get a longer interval?
Base it on impact on conformity rather than spend, then let performance move suppliers between tiers against published criteria. Typical criteria for extending an interval are two consecutive cycles with no major findings, all corrective actions closed on time, and a scorecard above an agreed threshold. Publishing the criteria matters, because a supplier who can see how to earn a longer interval has a reason to engage with the current cycle.

How often should supplier business reviews happen?
Relationship reviews follow a different clock from compliance data collection. Quarterly suits strategic and development suppliers where there is a genuine agenda, annual suits most others, and low-impact suppliers rarely need one at all. Keeping this separate from your data collection cycle stops the review from becoming a form-filling exercise.


Sources: Nadcap merit process and accreditation intervals, Thermal Processing Magazine; Performance Review Institute, frequently asked questions; Regulation (EU) 2017/625 on official controls, EUR-Lex; Background on the FDA Food Safety Modernization Act, FDA.

Every supplier programme eventually runs into the same wall, and it is rarely the wall people plan for.

The scorecard design is sound. The KPI weightings survived the workshop. The regulatory requirement is understood, the questionnaire is built, the deadline is in the calendar. Then the requests go out, and six weeks later 58 percent of suppliers have responded. Now the team is making decisions on partial data, and the programme’s credibility depends on a number nobody owns.

Response rate is the quiet constraint on supplier evaluation, supplier compliance data, onboarding and risk monitoring alike. It is also one of the most improvable numbers in procurement, because decades of research exist on what makes people answer a request for information, and very little of that research has made its way into how supplier data requests are actually written and sent.

This piece is the practical version of that evidence. The studies below come from survey methodology rather than procurement, so treat the effect sizes as direction rather than promises. The direction is consistent enough to design around.

Why the number matters more than it looks

A response rate is not just a completeness statistic. It quietly determines four things.

It sets the ceiling on your analysis. A supplier scorecard covering 60 percent of your base is not 60 percent of a scorecard programme. It is a programme with a blind spot in an unknown place, because non-responders are rarely a random sample. They tend to be the smaller, busier, less well-resourced suppliers, which is often exactly where the interesting variance sits.

It decides whether the programme survives its first year. Teams abandon supplier evaluation cycles because chasing felt worse than the insight was worth. High response rates make the next cycle easy to justify.

It is the real deadline on compliance work. Regulatory data collection depends on people outside your organisation finding time to answer you. The legal analysis takes a week. Getting documents back from 300 suppliers across nine countries takes months, which is why the ESG, CSRD and CSDDD requirements you collect during onboarding are a response rate problem long before they are a reporting problem.

It is a relationship signal. Suppliers who answer promptly are usually suppliers who feel the relationship is worth the effort. A rising response rate is one of the earliest positive indicators you get.

The good news is that high rates are demonstrably achievable at scale. CDP’s supply chain programme had more than 45,000 suppliers requested to disclose in 2025 through 200-plus buying organisations, and it publishes member examples at the top of the range, including Lenovo reaching a 98 percent response rate among its suppliers in 2024. That is not a typical number, but it does establish that the ceiling is much higher than most programmes assume.

What the research actually found

Three bodies of evidence are worth knowing.

A Cochrane methodology review by Edwards and colleagues pooled 32 trials covering 27 different strategies for electronic questionnaires and reported the odds of response for each. A meta-analysis by Wu, Zhao and Fils-Aime of 1,071 response rates drawn from 1,043 studies, in education-related research, found a weighted mean online response rate of 44.1 percent, with useful detail on what shifted it. And a controlled length experiment at the University of Gothenburg’s SOM Institute tested a long questionnaire against a shorter one on the same population.

Here is what moved the needle, and by how much.

What was tested Effect on odds of response Source
Shorter questionnaire Odds 1.73 times higher (95% CI 1.40 to 2.13) Cochrane review, electronic questionnaires
Topic framed as interesting or relevant Odds 1.85 times higher (1.52 to 2.26) Cochrane review
Telling recipients that others had responded Odds 1.52 times higher (1.36 to 1.70) Cochrane review
Offering to share the results back Odds 1.36 times higher (1.15 to 1.61) Cochrane review
Personalised request Odds 1.24 times higher (1.17 to 1.32) Cochrane review
Giving a deadline Odds 1.18 times higher (1.03 to 1.34) Cochrane review
Non-monetary incentive Odds 1.72 times higher (1.09 to 2.72) Cochrane review
Using the word “survey” in the subject line Odds 0.81, meaning response fell (0.67 to 0.97) Cochrane review
Contacting people before sending the request 54.6 percent responded versus 45.4 percent without pre-contact Wu and colleagues, meta-analysis
Following up by phone rather than email 79.0 percent versus 37.8 percent for email reminders, though based on only six cases Wu and colleagues, meta-analysis
Financial incentives No significant effect (45.6 percent versus 44.1 percent, p = .362) Wu and colleagues, meta-analysis
Longer questionnaire, 205 questions versus 149 45.8 percent versus 48.5 percent, a 2.7 point decline, in paper mode only Gothenburg SOM Institute length experiment

Two patterns jump out.

The first is that length and framing beat pressure. Shortening the request and making its relevance obvious carry more weight than deadlines or money. The Gothenburg experiment adds a useful nuance: the length penalty appeared in paper responses and not in web responses, which suggests that on a well-built digital form, perceived effort matters more than the raw question count. A long form that shows progress and remembers previous answers is treated differently from a long form that looks long.

The second is that the cheapest wins are in the wording. Personalisation, a named sender, a clear deadline, social proof and an offer to share results back cost nothing to implement and compound with each other.

Seven levers, in the order we would pull them

1. Cut the request to what you will actually use

Before touching the send mechanics, delete questions. Every field should have a named consumer and a decision it feeds. In most supplier questionnaires, a meaningful share of fields exist because someone might want them one day, and they are paid for in response rate by everyone.

If different supplier groups need different depth, send different forms. A supplier passing you a single certificate should never see the questionnaire designed for your strategic partners. A simple risk-based segmentation or a Kraljic-style split is usually enough to define two or three tiers of depth.

2. Pre-notify, especially the suppliers who matter most

Pre-contact lifted response from 45.4 to 54.6 percent in the meta-analysis, roughly nine points for one extra touch. In a supplier context, this is your category manager mentioning the upcoming request on the next call, or a short note from a named person a week ahead.

It works because it converts a cold request into an expected one. For your top suppliers, a two-minute mention in an existing conversation is the highest-return activity on this list.

3. Write it as a business request, not a survey

The Cochrane finding that the word “survey” in a subject line reduced response is a small effect with a large lesson. A supplier data request is not market research. It is part of a commercial relationship, and the wording should say so.

Name the sender and make them a real person. Say what the data is for, what decision it feeds and what the supplier gets from it. Give a date. Add social proof where you have it, since telling recipients that others had already responded raised the odds of response by around half. A line noting that most suppliers in their category have already completed it is honest, easy to generate from your own completion data, and effective.

These are the same principles that make supplier communication management work generally: clarity about purpose, a named owner, and a visible consequence.

4. Offer something back, and then deliver it

Offering to share results raised the odds of response by about a third, and non-monetary incentives nearly doubled them, while cash-style incentives showed no significant effect at all. For supplier programmes this is close to ideal, because the thing suppliers most want is something you already have.

Send the scorecard back. Show them where they rank against the anonymised peer group. Suppliers are rarely told how their buyer sees them, and many genuinely want to know. Doing this once changes next cycle’s response rate, because the request is no longer a tax, it is the price of admission to a conversation about their own performance. A structured business review is the natural place for that exchange.

5. Automate the reminder cadence, then escalate by channel

Response arrives in waves, and most programmes stop chasing before the second wave. Reminders work, and their effect depends on channel: phone follow-up reached 79 percent against 37.8 percent for email reminders in the meta-analysis, although that figure rests on only six cases and should be read as a strong hint rather than a benchmark.

The practical design is a ladder. Automated reminders on a fixed cadence handle the bulk, so nobody spends their week chasing 300 suppliers by hand. Then human escalation, by phone and through the category manager, is reserved for the suppliers who matter and the ones still missing. Automated follow-up is what makes that ladder affordable, and it is the single change that most reliably moves a stalled campaign.

6. Remove every avoidable point of friction

Each obstacle between the request and the answer costs you responses: a login nobody can find, a form that loses work, a file the supplier has already sent twice, a question only available in a language they do not work in.

Three fixes cover most of it. Let suppliers see and update what you already hold rather than re-entering it, which is what a supplier portal with clean master data is for. Support the languages and legal entities your supply base actually has, since multi-entity and multi-language onboarding removes a barrier that is invisible from headquarters. And never ask twice for the same document.

The internal version matters too. Evaluation fatigue on your own side produces vaguer requests and weaker follow-up, so protecting your team’s attention protects the response rate.

7. Make response rate a tracked KPI with an owner

Most programmes measure supplier performance and forget to measure the programme. Put the response rate on the dashboard: overall, by category, by region, by supplier tier and by days since request. Watch the curve rather than the endpoint, since a campaign that is flat in week three will not rescue itself in week eight.

Treat a persistent non-responder the way you would treat any other open issue, with an owner and a date, which is the same discipline as a supplier corrective action. And keep the record, because an audit trail of who was asked, when, and how many times is what turns a chase into evidence of due diligence.

What good looks like

Rather than a benchmark, use these as design targets, tested against your own baseline.

  • First send, no reminders: many programmes land somewhere near the 44 percent weighted mean the meta-analysis found for online surveys in its field, which is a rough reference point rather than a procurement benchmark. If you are well above it, your relationships are doing real work.
  • After a full automated reminder cadence: a large share of the remaining willing respondents arrive here. This is where most of the available improvement sits, and it is almost entirely a systems question.
  • After human escalation on the suppliers that matter: strategic and high-risk suppliers should approach complete coverage, because for those relationships a non-response is a commercial conversation rather than a data problem.
  • Next cycle: the strongest predictor of response is what happened last time. Suppliers who received something useful in return answer faster.

The shift from an annual push to a continuous rhythm helps here too. When requests are smaller and more frequent, each one is less disruptive, which is a large part of the case for moving from annual reviews to continuous monitoring.

Where software earns its place

Nothing on this list is impossible in a spreadsheet and an inbox. What a system changes is the cost of doing it consistently, at scale, cycle after cycle.

EvaluationsHub is built around exactly this layer of the work: per-supplier completion tracking so the gap is visible rather than estimated, automated reminders that keep the curve climbing without anyone chasing by hand, tiered request forms so each supplier group answers only what it should, supplier-facing access in the languages and entities your base actually needs, scorecards you can send back as the thing suppliers get in return, and an audit trail that assembles itself while the team works. Because it sits across the supplier lifecycle, the data you collect lands beside performance, risk and corrective actions instead of in another silo, which is the practical end of replacing spreadsheets.

If you want to see what your own response curve looks like, the fastest route is a focused pilot on one category, or a demo with your supplier list in front of you.

Frequently asked questions

What is a good supplier response rate?
There is no published procurement-wide benchmark, so the useful comparison is your own baseline. For context, a meta-analysis of 1,043 studies in education-related research found a weighted mean online survey response rate of 44.1 percent, and CDP publishes supply chain member examples at the top of the range, including a 98 percent supplier response rate reported by Lenovo in 2024. Most programmes should target near-complete coverage for strategic and high-risk suppliers and accept a lower rate on the long tail.

How do I increase supplier response rates?
The evidence points to shortening the request, making its relevance and purpose explicit, pre-notifying suppliers before it arrives, personalising it with a named sender, giving a clear deadline, adding social proof, offering something back such as the scorecard, and running an automated reminder cadence with human escalation for the suppliers who matter most.

Do incentives improve supplier response rates?
Financial incentives showed no significant effect in the meta-analysis, while non-monetary incentives nearly doubled the odds of response in the Cochrane review. In a supplier context the most valuable non-monetary offer is usually information the buyer already holds, such as performance feedback and anonymised peer benchmarking.

How many reminders should I send?
Enough to catch the second and third waves, on a predictable cadence, with the schedule automated rather than manual. Reminder effect varies sharply by channel, with phone follow-up outperforming email in the available research, so the efficient pattern is automated reminders for everyone and human follow-up reserved for strategic, high-risk and still-missing suppliers.

Does a longer supplier questionnaire reduce responses?
Yes, though the mechanism matters. Shorter electronic questionnaires had 1.73 times the odds of response in the Cochrane review. A controlled experiment comparing a 205-question form with a 149-question form found a 2.7 point drop, but only for paper responses, with no significant difference on the web. Perceived effort appears to matter more than question count, so a well-designed digital form with saved progress is treated more kindly than its length suggests.

Who should own the supplier response rate?
Someone with a commercial relationship to the supplier, supported by the programme owner. Compliance reminders move data slowly, and category managers move it quickly, which is why cross-department collaboration tends to be the difference between a stalled campaign and a complete one.

Should response rate be a reported KPI?
Yes. Tracking it overall and by category, region, tier and days since request turns a vague sense of “suppliers are slow” into a curve you can act on while there is still time to act, and it makes the programme’s own performance visible alongside the suppliers’.


Sources: Edwards PJ and colleagues, Methods to increase response to postal and electronic questionnaires, Cochrane systematic review; Wu M-J, Zhao K and Fils-Aime F, Response rates of online surveys in published research: a meta-analysis, Computers in Human Behavior Reports, 2022; Sandelin F, The effects of questionnaire length on response rate, non-response bias and data quality, SOM Institute, University of Gothenburg, 2022; CDP Supply Chain.

Just under 100 days separate us from 30 December 2026, the date the EU Deforestation Regulation starts to apply to large and medium operators and traders. Micro and small operators that are new to the rules have until 30 June 2027, although those already covered by the old EU Timber Regulation are in the December group.

After two postponements, this date looks firm. In its simplification review of May 2026 the European Commission confirmed that the regulation will not be reopened and that the existing timelines continue to apply. The July 2026 package then delivered the practical pieces teams had been waiting for: an updated product scope, a working information system, and guidance and FAQs in all EU languages.

Here is the encouraging part. Most of what EUDR asks for is supplier data, and collecting supplier data at scale is something procurement teams already know how to do. The teams that will be comfortable in December are not the ones with the best legal memo. They are the ones who started the data campaign early and kept the response rate high.

The short version

  • 30 December 2026 is the application date for large and medium operators and traders, plus micro and small operators previously covered by the EU Timber Regulation. 30 June 2027 applies to other micro and small operators.
  • Seven commodities are in scope: cattle, cocoa, coffee, oil palm, rubber, soya and wood, along with many derived products such as leather, chocolate, tyres, furniture and paper.
  • The core deliverable is a due diligence statement (DDS) submitted through the EU information system, which returns a reference number that travels down the chain.
  • Geolocation is the hard part. You need coordinates for every plot of land where the commodity was produced, plus evidence that production was legal in the country of origin.
  • Records are kept for five years, so this is a permanent data set, not a one-off filing.
  • Several simplifications genuinely reduce the workload, including one statement covering multiple shipments in defined cases, reuse of upstream reference numbers by downstream operators, and a group authorised representative.
  • Penalties are meaningful, with maximum fines of at least 4 percent of total annual EU-wide turnover, so this is worth doing properly.

What actually decides whether you are ready

Almost every EUDR article focuses on the legal obligations. That framing is useful for the general counsel and much less useful for the person who has to make it happen.

In practice, readiness comes down to one number: the share of your in-scope suppliers who have returned complete, usable data. A legal analysis can be finished in a week. Getting plot coordinates out of several hundred suppliers across multiple countries and languages takes months, because it depends on people outside your organisation finding time to answer you.

That reframing changes what you do first. Instead of starting with the regulation, start with the list of suppliers you need to hear from, and treat the next 14 weeks as a structured data campaign with owners, reminders and a visible completion rate. This is the same discipline as a good supplier onboarding programme, applied to a narrower question.

The data you need from suppliers

The regulation is specific about the information an operator has to collect and keep. For each in-scope product, that means:

  1. Product description, including HS code, trade name and the commodities it contains.
  2. Quantity, expressed in net mass and, where relevant, volume or number of units.
  3. Country of production, and for some chains the specific region.
  4. Geolocation of every plot of land where the commodity was produced. For cattle, this extends to every establishment where the animals were kept from birth to slaughter.
  5. Evidence of legal production under the laws of the country of origin, covering land use rights, environmental and forest rules, labour rights and tax and customs obligations.
  6. Supplier and customer details for products that already carry due diligence, including the reference numbers of the statements attached to them.

Two practical notes make this much easier to run. First, composite products need geolocation for each in-scope commodity they contain, so a chocolate bar with cocoa and palm oil generates two data trails rather than one. Second, the five-year retention requirement means this belongs in your permanent supplier master data, not in an inbox folder.

Where the simplifications genuinely help

The December 2025 revision and the 2026 packages removed real work. Four changes are worth building your plan around.

One statement can cover several shipments. Where the material comes from the same assessed harvest events and nothing in the supply chain has changed, a single due diligence statement can cover multiple shipments within a year. That turns a per-consignment task into a per-supply-chain task, which is a very different workload.

Downstream operators reuse reference numbers. Companies handling products for which due diligence has already been done collect and retain the upstream reference numbers rather than repeating the exercise. First-level downstream operators are not required to chase suppliers for those numbers either, since upstream operators provide them. If most of your volume is bought inside the EU from operators who have already filed, your job is largely to capture and store numbers reliably.

Groups can use one authorised representative. A single EU-established entity can submit statements on behalf of group members, which helps organisations with many legal entities. Legal responsibility still sits with each operator, so internal ownership still needs to be clear. If you run a multi-entity supply base, the same logic as multi-entity, multi-language onboarding applies here.

Country benchmarking reduces effort on low-risk origins. Where a commodity comes from a country classified as low risk, simplified due diligence applies and a full risk assessment is not required unless a substantiated concern arises. Knowing your origin mix early tells you where the real work sits.

Scope has narrowed in useful places. Printed products such as books and newspapers came out in December 2025. The July 2026 delegated act removed cattle hides and leather, retreaded tyres, soybeans for sowing, rubber articles, belts and vehicle seats, and added soluble coffee, certain palm oil derivatives and frozen cattle tongues, with the newly added products applying from 30 December 2027. Re-checking your HS codes against the current list is one of the highest-return hours you can spend this month.

A 14-week supplier data plan

Weeks 1 to 3: scope and segment

  1. Pull every purchased item against the current in-scope HS code list and confirm your own status as operator, trader or downstream actor. Scope errors are the most expensive mistake available, and they are cheap to fix now.
  2. Split in-scope suppliers into three groups: EU suppliers who will pass you a reference number, direct importers where you own the full due diligence, and suppliers whose status you cannot yet tell.
  3. Rank by volume and by origin risk. A simple risk-based segmentation tells you which twenty suppliers deserve a phone call rather than an email.

Weeks 4 to 7: launch the data request

  1. Build one request form per supplier group rather than one for everybody. A supplier passing you a reference number should not be asked for plot coordinates.
  2. Send it through a channel that tracks who has responded. A supplier portal with a visible completion status beats an email campaign, because you can see the gap rather than guess at it.
  3. Tell suppliers why you are asking and what happens if the data is late. Suppliers respond faster to a clear commercial consequence and a named contact than to a forwarded regulation.
  4. Set automated follow-up on a fixed cadence. Response rates rise sharply with the second and third reminder, and nobody has time to chase 300 suppliers by hand.

Weeks 8 to 11: assess, verify and close gaps

  1. Run risk assessment on what has arrived: check coordinates resolve to plausible plots, check legality evidence is current, and check the country classification you assumed.
  2. Escalate non-responders commercially. Category managers move data faster than compliance reminders do.
  3. For high-volume or high-risk origins, decide whether you need mitigation such as satellite checks, third-party verification or a remote supplier audit.
  4. Treat every missing or failed data point as a tracked action with an owner and a date, the same way you would handle a supplier corrective action.

Weeks 12 to 14: file and prove

  1. Register in the information system, dry-run a due diligence statement for one straightforward chain, and confirm the reference number flows to the people who need it.
  2. Decide which chains qualify for a single multi-shipment statement and document why, since that reasoning is what you will be asked about later.
  3. Make sure every decision, document and reference number sits in a reporting and audit trail that someone can navigate in minutes, with retention set to five years.

If you finish the first two blocks and are still collecting in week 12, that is normal. The value of starting now is that the gap is visible while there is still time to close it.

The data is worth more than the filing

It is tempting to treat this as a compliance cost. The more useful view is that you are about to build something your organisation has wanted for years: a verified map of where your material actually comes from, down to the plot.

That map answers questions well beyond deforestation. It tells you your true origin concentration, which is a supply continuity question. It feeds CSRD and CSDDD reporting rather than duplicating it, since much of the same evidence is requested twice under different names. It gives customers a credible answer when they flow their own requirements down to you. And it supports the wider supplier ESG picture you are already being asked for.

Teams that build this once, in a structured system, spend the following years maintaining it. Teams that build it in a spreadsheet for December rebuild it every time someone asks a new question. The same logic applies here as in the shift from annual reviews to continuous monitoring: the cadence you design now decides how much effort next year costs.

Where software helps

No platform makes you EUDR compliant, and any vendor who says otherwise is overselling. What good software does is remove the two things that actually break these programmes: chasing people, and losing evidence.

EvaluationsHub handles the supplier-facing layer of exactly this work. Structured data requests with per-supplier completion tracking, automated reminders that keep the response rate climbing, documents and evidence stored against the supplier record rather than in an inbox, and an audit trail that builds itself as the team works. It connects to the rest of the supplier lifecycle too, so origin and compliance data sits beside supplier scorecards, risk and corrective actions instead of in a separate silo. If you are also working through the broader due diligence agenda, our purchasing manager’s guide to CSDDD covers the neighbouring requirement.

If you want to see what this looks like with your own supplier list, you can run a focused pilot or book a demo.

Frequently asked questions

When does EUDR apply?
The EU Deforestation Regulation applies from 30 December 2026 for large and medium operators and traders, and for micro and small operators that were already covered by the EU Timber Regulation. Other micro and small operators have until 30 June 2027. Products added by the July 2026 delegated act apply from 30 December 2027.

Which products are covered by EUDR?
Seven commodities are in scope: cattle, cocoa, coffee, oil palm, rubber, soya and wood, together with many derived products including leather, chocolate, tyres, furniture and paper. The list has changed more than once, so check current HS codes rather than an older summary.

What supplier data does EUDR require?
Product description and HS code, quantity, country of production, geolocation coordinates for every plot of land where the commodity was produced, evidence that production was legal in the country of origin, and supplier and customer details including reference numbers of any existing due diligence statements. Records are kept for five years.

Do I need a due diligence statement for every shipment?
Not always. A single statement can cover several shipments within a year where the material comes from the same assessed harvest events and the supply chain has not changed. Once different harvest events are involved, a new statement is needed.

What if I only buy from EU suppliers?
You are likely a downstream actor for much of that volume. In that case your main obligation is to collect and retain the reference numbers of the due diligence statements provided by your suppliers, rather than repeating the underlying due diligence yourself.

What are the penalties for non-compliance?
Member States set penalties, with maximum fines for legal entities of at least 4 percent of total annual EU-wide turnover. Other sanctions include confiscation of products and revenues, temporary exclusion from public procurement and public funding, and a temporary ban on placing or exporting the products concerned.

Is EUDR going to be delayed again?
In its simplification review of May 2026 the Commission confirmed that the regulation will not be reopened and that existing timelines continue to apply. Planning around 30 December 2026 is the prudent assumption.


Sources: European Commission, Regulation on deforestation-free products; European Commission, Commission updates product scope and tools to support EUDR, 13 July 2026; Council of the EU, Council signs off targeted revision to simplify and postpone the regulation, 18 December 2025; Baker McKenzie, EU Commission publishes simplification review of EUDR, May 2026; Schoenherr, EU Deforestation Regulation: is your supply chain ready?.

On 16 September 2026, ISO published ISO 9001:2026, the first full revision of the world’s most widely used quality management standard in more than a decade. Certification bodies are already publishing gap guides, and most commentary focuses on leadership, quality culture and climate.

For supplier quality and vendor management teams, the more useful question is simpler: what does the new edition change about how we select, monitor and develop external providers, and what should we do about it this quarter?

The good news: nothing in ISO 9001:2026 asks you to rebuild your supplier programme. The revision rewards teams that already manage suppliers with structure, evidence and follow-through, and it gives everyone else a clear reason to get there over the transition period.

The short version

  • Clause 8.4 (control of externally provided processes, products and services) keeps its core logic. You still need defined criteria for evaluating, selecting, monitoring and re-evaluating external providers, with documented evidence.
  • Clause 6.1 now separates risks from opportunities, each with its own action cycle. That maps neatly onto supplier risk management on one side and supplier development on the other.
  • Clause 4.1 asks you to determine whether climate change is a relevant issue, which for most manufacturers quickly leads to the supply base.
  • Clause 4.2 asks which interested-party requirements you will actually address through the QMS, which matters when customers flow down supplier requirements.
  • Quality culture and ethical behaviour are now explicit in leadership and awareness requirements, and suppliers are part of how that culture shows up in practice.
  • A three-year transition is widely expected, in line with earlier revisions. Confirm exact dates with your certification body.

Why 8.4 staying stable is actually good news

Clause 8.4 has always been where supplier management lives inside ISO 9001. It requires an organisation to determine the controls it applies to external providers, and to base them on the potential impact of what those providers deliver.

The 2026 edition keeps that foundation. Clause 8.1 is aligned with 8.4 terminology so operational planning now explicitly covers externally provided processes, products or services that are relevant to the QMS. In other words, suppliers are part of the operation you plan, not a separate topic you audit once a year.

That stability gives supplier quality teams a practical advantage: the work you invest now in consistent evaluation criteria, performance data and follow-up will carry straight into your transition audit. If you want a refresher on what good looks like, our guide on how to build a weighted supplier scorecard covers the mechanics.

Where the 2026 revision touches your supplier programme

1. Risks and opportunities now run on separate tracks (6.1)

In ISO 9001:2015, risks and opportunities sat together, and in many QMS documents “opportunities” became an afterthought. The 2026 edition splits them: one action cycle for risks, one for opportunities. A note also links risk actions to maintaining conformity during and after disruptions.

For a supplier programme, this is a helpful framing:

  • The risk track covers single-source exposure, financial fragility, capacity, quality escapes, geopolitical and logistics disruption. This is the work of supplier risk management: identifying which suppliers could interrupt conformity, and what you do before it happens.
  • The opportunity track covers suppliers who could improve cost, quality, innovation or sustainability if you invest in the relationship. This is where supplier development programmes and structured supplier performance improvement belong.

An auditor asking “how do you address opportunities?” is a chance to show that your best suppliers are managed actively, not only your worst ones.

2. Disruption resilience becomes part of the quality conversation

The link between risk actions and conformity during disruption reflects what quality teams have experienced over the past few years. It is hard to demonstrate that link with an annual supplier survey and a spreadsheet.

What helps is a monitoring rhythm that surfaces early signals: late deliveries trending up, a drop in responsiveness, repeated minor nonconformities. Moving from annual reviews to continuous supplier monitoring is one of the most practical steps you can take here, and it pays off in daily operations long before the audit.

3. Climate change relevance (4.1) reaches the supply base

The 2024 climate amendment is now fully integrated: organisations must determine whether climate change is a relevant issue for their context. For most manufacturers, a large share of climate exposure sits with suppliers, whether as physical risk to supplier sites or as emissions data customers are starting to request.

ISO 9001:2026 does not turn your QMS into a sustainability report. But if you conclude climate is relevant, your supplier evaluation criteria are a natural place to reflect it. Teams already collecting supplier ESG and CSRD data can connect those two worlds rather than running them in parallel.

4. Interested-party requirements you commit to (4.2)

The new 4.2 requirement asks you to determine which interested-party requirements will be addressed through the QMS. In practice, a lot of those arrive as customer flow-downs: supplier approval rules, special characteristics, traceability, audit rights.

Being explicit about which of these you manage through your supplier processes makes audits smoother and makes the conversation with customers easier. For automotive suppliers, our article on building a supplier scorecard that satisfies IATF 16949 shows how customer-specific requirements translate into scorecard criteria.

5. Quality culture and ethics include how you treat suppliers

Leadership must now demonstrate commitment to quality culture and ethical behaviour, and people must be aware of what that means (7.3). Culture is visible in behaviour, and supplier management is one of the most visible places: whether evaluations are fair and consistent, whether feedback reaches the supplier, whether corrective actions are closed or quietly forgotten.

Consistency across evaluators and sites matters here. Our piece on reducing bias in supplier performance reviews offers a few simple design choices that make evaluations more defensible and more useful for the supplier.

A practical 90-day plan for supplier quality teams

You have time. The aim of the first 90 days is not to be “2026-certified” but to know where you stand and to start collecting the evidence you will need anyway.

Days 1 to 30: map and baseline

  1. Get the standard and your certification body’s transition guidance, and note the dates that apply to you.
  2. List your current supplier evaluation, selection, monitoring and re-evaluation criteria (8.4.1) and check they are applied consistently across sites and categories.
  3. Segment suppliers by impact on conformity. A simple Kraljic-based segmentation is enough to decide where controls should be tighter.

Days 31 to 60: split risks and opportunities

  1. For critical suppliers, record the top risks to conformity and the planned actions, including continuity during disruption.
  2. Separately, pick three to five suppliers with clear improvement or innovation potential and define what you want to achieve with each.
  3. Decide whether climate change is relevant for your context, and if so, which supplier criteria will reflect it.

Days 61 to 90: close the loop and prove it

  1. Make sure every nonconformity from a supplier leads to a tracked supplier CAPA with an owner, a due date and verification of effectiveness.
  2. Check that evaluation results, decisions and follow-ups are retained as documented information that an auditor can navigate in minutes. A clear reporting and audit trail is the difference between a stressful audit and a calm one.
  3. Share results with suppliers. Feedback that never reaches the supplier improves nothing.

Where software helps (and where it does not)

A standard cannot be met by a tool, and ISO 9001:2026 is deliberately technology-neutral. What software does well is make good practice the default: the same criteria applied by every evaluator, reminders that keep reviews on cadence, CAPAs that cannot silently stall, and an audit trail that builds itself as you work.

EvaluationsHub is built for exactly this layer of supplier lifecycle management, from supplier onboarding and supplier scorecards to risk, CAPA and development. If you want to see how your current process would look in a structured setup, you can start a supplier performance pilot or book a demo.

Frequently asked questions

When was ISO 9001:2026 published?
ISO published ISO 9001:2026 on 16 September 2026. It replaces ISO 9001:2015.

How long is the ISO 9001:2026 transition period?
A three-year transition period is widely expected, in line with previous revisions, which points to late 2029. Your certification body will confirm the exact dates and audit rules that apply to you.

Does ISO 9001:2026 change clause 8.4 on external providers?
The core requirements of 8.4 remain: defined criteria for evaluation, selection, performance monitoring and re-evaluation of external providers, with documented evidence. The wider revision (separate risk and opportunity cycles in 6.1, climate relevance in 4.1, interested-party requirements in 4.2, quality culture and ethics) influences how you apply those controls.

Do we need new supplier evaluation criteria for ISO 9001:2026?
Not necessarily. Many organisations can keep their criteria and strengthen consistency, risk and opportunity actions, and evidence. If you decide climate change is relevant to your context, reviewing supplier criteria with that lens is a sensible step.

Is a supplier scorecard enough to meet clause 8.4?
A scorecard covers performance monitoring, which is a central part of 8.4. Auditors will also look at selection criteria, re-evaluation, actions on poor performance and documented evidence, so the scorecard works best as part of a connected process.


Sources: ISO news release, September 2026; ISO 9001:2026 standard page; CQI | IRCA revision guidance; TÜV Rheinland revision overview; SGS publication note.

Published by EvaluationsHub. We make one of the products discussed below, so we have described every other vendor using their own public pages and linked each claim. Vendor information last reviewed 29 September 2026.

Every supplier management vendor now has an AI story. Some call it an assistant, some a copilot, some a fleet of agents. The demos look alike: a chat box, a summary of a supplier, a risk alert that appears on cue. The differences that matter for a supplier quality or vendor management team sit underneath, and they rarely show up in a standard demo script.

This guide gives you a way to compare them. It is written for people who have to live with the result: the supplier quality engineer who owns the CAPA, the category manager who owns the QBR, and the procurement lead who has to explain an AI decision to an auditor.

Why comparing AI features is harder than it looks

Buyers are clearly interested and still cautious. In a Gartner survey of 101 chief procurement officers run in January and February 2026, only 36% said they were very confident in their ability to redesign roles and processes around AI. The Hackett Group’s 2026 Procurement Key Issues study found that 43% of organisations are actively pursuing AI deployment, but only 12% report large-scale implementation, and that 69% access AI through capabilities embedded in their existing procurement platforms.

That last figure is the important one. Most teams will not buy AI on its own. They will get whatever AI comes with the supplier platform they choose. So the AI question is really a platform question, and it deserves the same rigour as the rest of your evaluation.

Three things make it hard:

  • The same words mean different things. “Agent” can mean a model that drafts an email or one that changes supplier master data without asking.
  • Availability varies. Several capabilities announced in 2025 and 2026 are in beta, rolling out, or planned for a later release. Vendors usually say so on their own pages, but not in the demo.
  • AI is only as good as the supplier data under it. A brilliant model reading a stale supplier record produces a confident, wrong answer.

Compare the loop, not the model

Almost every vendor builds on one of a handful of large language models. Comparing model names tells you little. What differs is the loop the AI sits in. For supplier management, that loop has five steps, and you can compare every vendor on each one:

  1. Read. Which of your data can the AI actually see? Scorecards, contracts, CAPA history, supplier portal submissions, external risk feeds, ERP transactions?
  2. Reason. Does it show why it reached a conclusion, with the data points and sources behind it?
  3. Recommend. What does it propose: a summary, a score, a corrective action, a sourcing event, a message to a supplier?
  4. Act. What happens next? Does a person approve, edit or dismiss, or does the AI proceed on its own? Which actions can it never take alone?
  5. Record. Is every AI output, approval and outcome logged in a way you could export and hand to an auditor?

A capability that is strong at step 3 and silent on steps 4 and 5 is a nice demo. A capability that is solid across all five is something you can put into a regulated supplier quality process.

Eight questions to ask every vendor

Use these in the RFP and again, live, in the demo. They are deliberately practical.

1. What supplier data does the AI read, and how fresh is it?

Ask for a list of the objects the AI can access. The most useful supplier AI connects things that normally live apart: a contractual KPI in one place and the measured OTIF or PPM in another. If the AI only reads documents you upload into a chat, it is a writing aid, which is useful but different.

2. Can you show me the reasoning behind one recommendation?

Pick a recommendation in the demo and ask to see exactly which data triggered it. You want to see the threshold, the measured values and the time window, not a paragraph of fluent prose.

3. What can it do without a human, and can I switch that off?

This is the single most important question. Ask for the list of actions the AI can take autonomously, the list it can only propose, and whether you can configure the boundary per action type. Supplier-facing communications, awards and changes to bank details deserve particular attention.

4. Where is the audit trail, and can I export it?

For teams working under ISO 9001, IATF 16949, GMP or medical device rules, an AI-proposed corrective action is still a corrective action. You need to show who approved it, when, and on what evidence.

5. Is my data used to train models, and where is it processed?

Get the answer in writing, in the contract, not just on a web page. Ask separately about the vendor’s own models and any third-party model providers they use.

6. What is generally available today, and what is on the roadmap?

Ask the vendor to mark each capability you saw as generally available, beta, or planned, with dates. There is nothing wrong with a roadmap. There is something wrong with buying one by accident.

7. What happens when the AI is wrong?

Ask to see a dismissed recommendation. Does the dismissal get recorded? Does the system learn from corrections? Can you tune thresholds so the queue does not fill with noise?

8. Can you run it on our data before we sign?

A demo on the vendor’s sample data proves the interface works. A short session on a slice of your own supplier data proves the AI works for you.

What leading vendors say their AI does

The table below summarises how each vendor describes its own AI for supplier management, using its public pages as of 29 September 2026. It is a starting point for your own questions, not a ranking, and features change quickly. Where a vendor’s page did not address a topic, we say “not stated on the page reviewed”. That does not mean the capability is absent, only that you should ask.

Supplier management AI as described by each vendor (public sources, reviewed 29 September 2026)
Vendor AI layer Supplier management examples, in the vendor’s framing Oversight and data statements published A genuine strength
SAP (Ariba) Joule, Supplier Management Assistant Sanctions screening, financial health checks and risk evaluations; supplier classification; onboarding data enrichment; “corrective action plans before performance risk materializes” Not stated on the supplier management page reviewed; see SAP’s general AI ethics principles Built on a long-established supplier lifecycle module and third-party data enrichment
Coupa Navi agents Supplier Discovery Agent and Supplier Assistance Agent (announced November 2025) Not stated in the announcement reviewed Community data from a very large buyer and supplier network
Ivalua IVA Validating documents and enriching profiles at onboarding; keeping risk and performance scores current “Every AI action has a person accountable”; continuous audit trails; IVA inherits user permissions; “never uses your data to train LLMs” Among the clearest published governance commitments
JAGGAER JAI Supplier self-service onboarding and cited answers across procurement data “JAI suggests. You decide. Always.” No customer data used for model training; data processed in the customer’s selected region Explicit data residency statement and source citations on answers
Oracle Fusion Procurement Supplier Qualification Workspace (agentic app) AI summary of qualification compliance; flags missing or adverse qualifications; currently up to 500 strategic suppliers across three tiers “Review AI recommended priority actions for accuracy before applying” Precise release documentation, including stated limits
Gatekeeper Gatekeeper AI and agents SLA Compliance Check Agent; Bank Details Validator Agent; claims 50+ agents Not stated in detail on the page reviewed A wide catalogue of narrow agents on one vendor and contract record
Kodiak Hub AI across the platform AI-enhanced insights for supplier performance, risk forecasting and anomaly detection; document extraction ISO 27001 stated; AI oversight not stated on the page reviewed AI inside a full SRM loop of ratings, goals and actions
Prewave AI-based risk alerts Prioritising risks by AI-driven impact, probability and relevance; deep-tier mapping Not stated on the page reviewed External risk sensing across languages and supply tiers

Two patterns stand out. First, governance disclosure is uneven. A few vendors publish specific commitments on human accountability, training data and residency; many do not address them on their AI pages at all. Second, most published examples cluster around risk sensing, onboarding documents and sourcing. Fewer vendors describe AI that connects what a supplier promised in a contract with how that supplier is actually performing, which is where supplier quality teams spend most of their week.

Suites, specialists and risk platforms: which kind of AI fits you

A source-to-pay suite’s AI is a strong fit if you already run that suite, your priority is transactional efficiency across sourcing, contracts and payables, and you have an enterprise agreement in place. The AI benefits from seeing spend and transaction data in one system.

A risk intelligence platform’s AI is a strong fit if your main exposure is external: disruption, sanctions, adverse media or deep-tier visibility. It watches the world outside your supply base very well.

A supplier performance and lifecycle specialist’s AI is a strong fit if your pain sits in the relationship itself: scorecards, corrective actions, QBRs, contractual KPIs, supplier development, and the evidence trail auditors ask for. Here the AI can be very close to the data that quality and vendor management teams work with every day.

Many large organisations will combine a suite with a specialist. The limits of S2P suites for supplier relationship work and the ERP versus specialised SPM question are covered in separate articles.

Where EvaluationsHub, InitiativesHub and Eva AI fit

We are a specialist in supplier performance, and we built our AI around the loop described above rather than around a chat window. Three products work on one supplier data model:

  • EvaluationsHub covers supplier performance and risk: weighted supplier scorecards per segment, risk and ESG data collection, corrective actions (CAPA), QBR management and a two-way supplier portal.
  • InitiativesHub covers execution: intake, RFx, supplier qualification, awarding, contract management with SLAs and KPIs mapped to clauses, and onboarding that hands over to EvaluationsHub when it is complete. Together, the two cover the full supplier lifecycle.
  • Eva AI is the AI layer on top. It is designed to do three things.

Contract and performance cross-check. Eva reads the KPI obligations in your contracts and compares them with live scorecard data. When a deviation crosses a threshold you define, it proposes a corrective action with its reasoning attached. For example: a contract requires 97% OTIF, the 30-day average is 88.4%, and the gap has exceeded your 5-point threshold for three consecutive weeks.

Risk and compliance monitoring. Eva watches risk signals across the supplier portfolio against your acceptance frame and proposes the right compliance workflow, linked to the relevant contract clause.

Cost reduction intelligence. Eva looks for consolidation opportunities, suppliers ready for renegotiation based on performance, and pricing anomalies, and proposes structured cost reduction projects in InitiativesHub.

On the questions above, our answers are simple. Eva recommends and you decide: every action is held for human review, with approve, edit or dismiss in one click. No supplier-facing action happens without explicit approval. Confidence thresholds are configurable per action type. Every AI recommendation, approval and outcome goes into an immutable, exportable audit trail. The platform is ISO 27001 certified and hosted in the EU.

We are not the right choice for everything. If you want one vendor for requisitions, invoices and payments, a suite will serve you better. If deep-tier news monitoring in dozens of languages is your main need, a dedicated risk intelligence platform goes further. Our focus is the performance, risk and improvement loop with your suppliers, and the AI that makes that loop faster without taking the decision away from your team.

A note on regulation

Human oversight is becoming a regulatory expectation, not only good practice. Under the EU AI Act, high-risk AI systems must be designed so that people can oversee them effectively. In May 2026 the Council and Parliament agreed to move the application dates for high-risk obligations to 2 December 2027 for stand-alone high-risk systems and 2 August 2028 for those embedded in products. Most supplier management uses are unlikely to fall in the high-risk categories, but the oversight principle is a sensible benchmark for any AI that proposes actions affecting suppliers. Our EU AI Act guide for procurement goes further; check any specific obligation with your legal team.

What this comparison cannot tell you

We compared public statements, not live systems. We did not test any vendor’s AI on real data, pricing for AI add-ons is rarely published, and some capabilities may be further along or further behind than a web page suggests. Ask each vendor, including us, to show you the loop working on your own suppliers.

Frequently asked questions

What is the most important AI feature in SRM software?

The ability to connect data that normally sits apart, such as contract obligations and measured performance, and to turn a deviation into a proposed action with clear reasoning. Summaries and chat are helpful, but the value is in the action loop.

Should AI in supplier management act without human approval?

For low-risk, reversible, internal tasks, limited autonomy can be reasonable. For supplier-facing communications, awards, corrective actions and any change to supplier bank details, human approval with an audit trail is the safer design and the one auditors will expect.

How do I check whether a vendor’s AI feature is actually available?

Ask the vendor to label each capability as generally available, beta or planned, with dates, and put that list in the contract schedule. Many vendors state availability on their own release notes or press releases, so check those too.

Do I need clean supplier data before AI is useful?

You need reasonably current core data: supplier records, a scorecard with defined KPIs, and your key contract terms. You do not need perfection. A good platform will show you where data gaps are weakening its recommendations.

If you want to see the loop on your own supplier data, book a demo and bring one supplier you are worried about.

For a fuller treatment of what agents should and should not do, read our guides on AI agents in procurement and AI agents in supplier quality.

By Prof. Dr. Bert Paesbrugghe, founder of EvaluationsHub and associate professor at IÉSEG School of Management.

A kitchen can have a superb fishmonger and still run out of fish on a Saturday night. I learned that as a chef long before I studied buyers for a living. The fishmonger was financially sound, delivered great product and never once missed an invoice. The problem sat elsewhere: one boat, one road, one driver. When any of the three failed, the supplier was fine and my dining room was not.

Procurement is now discovering the same lesson at scale. ProcureAbility’s 2026 CPO Benchmark Study, published on 9 September 2026 and based on 160 senior procurement leaders at large enterprises across 21 industries, describes what it calls a shift in procurement’s risk focus from supplier health to supply continuity. Read that sentence twice. For years we asked whether the supplier is a sound business. Increasingly we are asking whether the thing we need will actually arrive.

The numbers behind that shift are worth sitting with. In the same study, 62% of respondents cite geopolitical disruption as a primary risk and 55% cite logistics disruption. Yet only 6% say they have predictive or real-time monitoring tools, while 62% still rely on manual or periodic risk assessment. So buyers worry about continuity more than anything, and most of them check on it the way a chef checks a walk-in fridge: by opening the door now and then and hoping.

Why this is a relationship question, not only a tooling question

When I interviewed buyers for my research, continuity came up again and again, often before price did. The sentence behind it was rarely “we want the cheapest supplier” and far more often “I never want to explain to my board why the line stopped”. Business continuity is the buyer’s quiet top priority, and it shapes how they judge every supplier interaction, whether or not they say so.

Here is the part that interests me most. The early warning signs of a continuity problem usually sit with the supplier long before they sit in any dashboard. A second-tier component maker is struggling. A port is congested. A key engineer has left. The supplier often knows first. Whether the buyer hears about it depends on something no monitoring tool can buy: the quality of the relationship.

That is where balance enters. If a supplier expects to be punished for bad news, bad news arrives late. If a buyer never explains which parts of the supply are truly critical, the supplier cannot prioritise them. Both sides are guessing, and the guessing happens in the space where continuity is won or lost.

Three habits that build continuity before the crisis

Say what “critical” means. Most buyers have a mental list of the three or four things that would stop the business. Few suppliers have ever seen that list. Share it. A supplier who knows that one item out of forty matters enormously will treat it differently on a bad week. This is mise en place for a relationship: everything in its place before service starts.

Make bad news cheap. Ask suppliers, in a regular conversation and not only in a quarterly review, what worries them about the next ninety days. Then respond to the answer with help, not a penalty clause. A supplier who is thanked for an early warning will send the next one earlier.

Swap the snapshot for a rhythm. A yearly risk assessment is a photograph of a river. Continuity moves. Even without expensive tooling, a light monthly check-in on the handful of critical suppliers gives you more signal than a thick annual questionnaire. Technology can then amplify that rhythm, but it should follow a logic you have already agreed on.

The balance point

There is a temptation, when risk headlines pile up, to tighten control: more audits, more clauses, more reporting demands. Some of that is sensible. Yet a supplier buried in compliance requests has less attention left for your actual supply, and the relationship becomes a paperwork exercise that produces reassuring documents and thin resilience.

The alternative is equilibrium. The buyer is clear about what matters and generous with context. The supplier is honest about risk and early with news. Neither side wins at the other’s expense, because the outcome both want, goods arriving on time, is shared.

I explore this buyer-side perspective, including what buyers told me they wish their suppliers understood, in my book The Buyer’s Balance: What your customers want to share with you (Owl Press, 2025, ISBN 9789464778908). More at evaluationshub.com/book.

So here is my question for this week. If your most critical supplier had a problem today, how many days would pass before you heard about it, and from whom?

Source: ProcureAbility, 2026 CPO Benchmark Study, press release of 9 September 2026 (160 senior procurement leaders, 21 industries).

Ask a purchasing manager what keeps them awake at night and you will rarely hear “unit price”. You hear a version of the same sentence, phrased a dozen different ways: the line has to keep running.

I have interviewed a lot of buyers over the past decade for my research, and this is the most consistent thing I have found. Savings get you a good annual review. A stoppage gets you a very different conversation with the board. Continuity is not one priority among several for most purchasers but the floor that everything else stands on.

So here is the uncomfortable question. If continuity is the thing you care about most, how much of your actual supplier conversation is about it?

What the numbers say

QIMA’s 2026 Global Sourcing Survey, which draws on the experiences of more than 1,000 businesses with international sourcing networks, found that the average company now maps 60% of its supplier network, up from 53%. That is real progress. And yet only 18% report full end-to-end visibility. Four out of five buying organisations are making continuity decisions with a partial map.

The same survey found that 79% of supply chains expect costs to be a major disruption in 2026, and that 43% shifted sourcing locations during 2025. The map is changing faster than most organisations can redraw it.

Sit with that for a moment. Your top priority is the thing you can see least clearly.

Your supplier can see what you cannot

Here is what I find strange about the way most buyer-seller conversations are structured. The person sitting across the table often knows exactly the part of the chain you have never mapped. They know which of their own inputs is single-sourced. They know their tier-two supplier has been quoting longer lead times since spring. They know the plant in question runs at 95% and has no slack for your rush order in November.

They know all of it. And in a great many relationships, nobody ever asks.

When I was still cooking, mise en place was not a nice-to-have. You did not discover at seven in the evening that the fish had not arrived. You found out at ten in the morning, because you had a relationship with the person who brought the fish, and that person told you things before you had to ask. That relationship was worth considerably more than the three cents a kilo you might have saved somewhere else.

Procurement has exactly the same dynamic and usually far less of the conversation.

Why buyers hold back

Buyers have reasons for not opening up about continuity risk. If you tell a supplier you have no alternative source, you have told them something useful about your negotiating position. If you tell them the launch date is immovable, you have handed them leverage.

I understand the instinct. I also think the arithmetic has changed. The cost of an information gap that hides a disruption is now much higher than the few points of margin that gap protects. A supplier who does not know that a particular component is critical to you will not prioritise it when their own capacity tightens, not out of obstruction but out of simple ignorance.

This is the balance I write about in The Buyer’s Balance: What your customers want to share with you (Owl Press, 2025). Equilibrium in a B2B relationship is not a matter of both parties being pleasant to each other but of both parties holding roughly the same picture of what is going on, so that each can act sensibly on it.

Three questions worth asking this quarter

If continuity is genuinely your top priority, put it into the conversation rather than into a clause.

Where is your own single point of failure? Not yours as in your company. Theirs. Most suppliers will answer honestly if you ask in a way that does not sound like the opening of an audit.

What would you need from us to hold our lead time if your capacity tightened? Usually the answer is a forecast, or earlier notice, or a commitment they can plan against. Often it costs you nothing at all.

What have you told other customers that you have not told us? Slightly cheeky. Very revealing.

None of these show up on a supplier scorecard. All three of them do more for continuity than another paragraph about liquidated damages.

The consistency point

A supplier who tells you about a problem three weeks early is worth more than one who never has problems on paper. Predictability beats brilliance. A vendor who performs at eight out of ten every single time is far easier to build a business around than one who swings between ten and four.

And if you want that from your suppliers, you have to be that for your suppliers. Consistent forecasts. Consistent feedback. Consistent honesty about what you actually need and when.

Your top priority deserves more than a paragraph in the contract.


Bert Paesbrugghe is associate professor of sales management and purchasing strategy at IESEG School of Management and author of The Buyer’s Balance: What your customers want to share with you (Owl Press, 2025), ISBN 9789464778908. More at evaluationshub.com/book.

Source: QIMA, 2026 Global Sourcing Survey: From Disruption to Opportunity (2026), based on responses from over 1,000 businesses with international sourcing networks. qima.com