AI Integrity and Transparency

Business Case

Brand reputation

If suppliers use AI in ways that produce harmful, deceptive or discriminatory outcomes in their own operations, the buyer's association with those suppliers creates reputational exposure.

Financial resilience

Without strict controls on how AI systems handle procurement data, confidential commercial information, supplier pricing and supply chain IP can leak into external systems, third-party model logs or training datasets, creating direct financial exposure through competitive harm, loss of negotiating advantage and breach of confidentiality obligations.

Informed Decision Making

Without clear AI governance structures, buyers cannot demonstrate that procurement decisions were made fairly or traceably, creating risk in the event of challenge, audit or regulatory scrutiny.

Legal and regulatory compliance

AI used in procurement to evaluate suppliers, score bids or flag risks can introduce bias, hallucination or errors, resulting in unfair treatment, disadvantage for particular supplier groups or even unlawful discrimination, exposing the buyer to legal risk. Buyers that establish responsible AI governance ahead of mandatory requirements are also better positioned for compliance when laws do come into place and are better able to influence the standards being set.

Productivity and quality

Introducing AI into the procurement function without a clear workforce policy risks displacing experienced procurement professionals whose judgment, relationships and contextual knowledge are essential to responsible sourcing, while also undermining staff trust in the organization. AI governance protects the integrity of procurement processes and controls, maintaining the fairness, competition and auditability on which sustainable procurement depends.

Stronger supplier relationships

Suppliers that cannot understand how AI-assisted procurement decisions were reached or that suspect AI is producing biased or opaque outcomes lose trust in the process, resulting in capable suppliers disengaging from the buyer.

Strategy & Planning

Assessment

Initial Action

In collaboration with IT colleagues, conduct a basic review on how AI is currently being used within the procurement function, such as:

  • whether AI is used for procurement decisions, including supplier scoring, bid evaluation, risk flagging, spend analysis and contract review
  • whether procurement teams understand how AI tools arrive at their recommendations
  • whether procurement colleagues review and approve AI recommendations for a procurement decision
  • whether suppliers are informed about AI use in the procurement process
  • whether there is any risks of information leakage from currently used AI tools
Intermediate Progress

In collaboration with IT colleagues, conduct a detailed review of AI use across procurement categories and processes, such as:

  • which AI tools are being used at each stage of the procurement lifecycle and what level of influence they have on decisions
  • whether the data used to train or inform AI tools avoids bias
  • whether current data quality and security practices are sufficient to support responsible AI use in procurement
  • whether any AI tools currently used in procurement were onboarded without a risk assessment
  • whether current documentation practices for decisions influenced by AI are sufficient
  • whether AI tools are properly serving their function or leading to mistakes / inconsistent output
Advanced Practice

Conduct a review of how strategic suppliers are using AI in their own operations and tender processes and whether supplier AI use introduces bias, opacity or integrity risks that could affect the buyer, such as:

  • whether suppliers are using AI to generate tender responses or sustainability reports submitted to the buyer, and whether human oversight is maintained to avoid misrepresentation or inaccuracies about the supplier's real capabilities
  • whether suppliers have a responsible AI use policy for their own operations and procurement practices to avoid discriminatory outcomes or systemic errors that create legal and reputational exposure for the buyer
  • whether suppliers have experienced AI-related incidents in the past three years

Policy

Initial Action

Adopt and train procurement staff on a responsible AI use policy for procurement, aligned with the OECD AI Principles, UNESCO Ethics of AI Recommendation or a similar framework, covering:

  • accountability: procurement colleagues must take ownership of every AI-assisted procurement decision, validating AI recommendations before they influence any decision
  • transparency: recommendations from AI tools used in procurement should be acted upon only if procurement staff can explain factors and data behind that conclusion
  • human oversight: AI may support but never replace human judgment in procurement decisions
  • supplier disclosure: suppliers must be informed where AI is used or has influenced a procurement decision affecting them
  • legal and compliance integration: any AI tool handling supplier or personal data must comply with applicable privacy law and procurement policies, as well as take the correct measures to prevent leakage
Intermediate Progress

Expand policy to cover:

  • workforce support: AI is introduced to support the procurement function rather than eliminating the professional judgment of responsible sourcing colleagues; if AI is introduced into the procurement function in ways that affect the roles or responsibilities of procurement staff, the company commits to managing that transition responsibly, including providing retraining opportunities and support for affected employees
  • escalation: a clear path must exist for procurement staff and suppliers to raise concerns about AI-assisted decisions
  • fairness: AI tools must be regularly checked for bias, with diverse and representative training data and rules set to prevent systematic exclusion of particular supplier groups and regular reviews to detect and address emerging fairness issues
  • AI assurance: Add Human-in-the-loop requirements and algorithm transparency audits as mandatory elements, requiring that all high-risk AI applications in procurement have documented human-in-the-loop checkpoints and are subject to periodic independent algorithm transparency audits
  • AI use restrictions as needed for high-risk scenarios
Advanced Practice

Require strategic suppliers to adopt their own responsible AI use policies covering their procurement and operational functions with clear standards for what the buyer expects from suppliers in terms of AI transparency, human oversight and bias prevention.

Targets

Initial Action

Set a target for 100% AI-assisted procurement decisions to be documented with a record of the human review applied.

Intermediate Progress

Set targets for the percentage of AI-assisted procurement processes that have been audited for bias and fairness, as well as targets for secure, approved AI platforms and decommissioning of unapproved AI tools.

Advanced Practice

Set targets for the percentage of strategic suppliers with responsible AI governance frameworks in place.

Supplier Selection

Screening

Initial Action

Review the buyer's own use of AI in the pre-qualification and RFP process for practices that may create unfair outcomes for suppliers, such as:

  • using AI to screen or score suppliers without disclosing that AI was involved or how it influenced the outcome
  • relying on AI tools trained on historical data that may reflect past procurement biases
  • using AI to generate or evaluate tender requirements in ways that may inadvertently favor incumbent suppliers
  • applying AI to risk evaluation or decisions without human review
  • not maintaining clear records of AI deployment, decision rationale and human oversight measures applied
Intermediate Progress

Include questions in RFPs to understand how suppliers are using AI in the tender process, covering:

  • does the supplier use AI to generate or support the preparation of tender responses, and, if so, how is human review and accountability maintained?
  • is the supplier using third‑party platforms with embedded AI tools?
  • what governance structures or policies does the supplier have in place to ensure AI is used transparently and without harmful bias or inaccuracies?
  • has the supplier experienced any AI-related incidents, errors or complaints in the past three years, and how were they managed?
Advanced Practice

Require strategic suppliers to provide evidence of their own responsible AI governance policies and practices as a condition of RFP eligibility in categories where AI is material to their operations or where supplier AI use creates significant exposure for the buyer.

Selection

Initial Action

Ensure that AI-assisted supplier evaluation is transparent and accountable, including requiring that every AI recommendation in supplier selection is reviewed and validated by a procurement professional before influencing the final decision and that the rationale for the decision can be explained to the supplier upon request.

Intermediate Progress

For AI generated tender responses, conduct additional due diligence to confirm the accuracy and integrity of the submission as part of the selection process.

Advanced Practice

Assess the quality, maturity and credibility of strategic suppliers' responsible AI governance as a weighted criterion in tender evaluations, including whether they have documented policies on responsible AI use, whether human oversight is maintained over AI-assisted decisions, records on AI related incidents, audit and security of AI tools and whether they cascade responsible AI practices to their own supply chains.

Contracts

Initial Action

Include provisions in all supplier contracts committing the buyer to informing suppliers when AI has materially influenced a procurement decision affecting them, and providing a mechanism through which suppliers can request an explanation of the outcome.

Intermediate Progress

Require suppliers to clearly identify where AI has been used in any documentation submitted to the buyer.

Advanced Practice

Require suppliers to flag use of AI that could affect the buyer's operations, data or exposure, as well as to notify the buyer of any AI-related incident in their operations that could create reputational or ethical risk for the buyer.

Supplier Engagement

Goal Setting

Initial Action

Work with suppliers that use AI in processes that affect the buyer to set goals for implementing basic responsible AI controls, including a documented AI use policy, human oversight over AI-assisted decisions and a process for disclosing AI use to the buyer within a defined timeframe.

Intermediate Progress

Work with strategic suppliers categories to set specific improvement goals on responsible AI use, including timelines for implementing AI governance policies, completing bias audits of AI tools used in their operations and establishing human oversight mechanisms for AI-assisted decisions that affect their own supply chains or the buyer's interests.

Advanced Practice

Co-develop multi-year responsible AI governance roadmaps with strategic suppliers, including measurable targets for strengthening AI governance frameworks, cascading responsible AI requirements through their own supply chains and continuous improvement targets informed by bias audit findings and incident data.

Capacity Building

Initial Action

Provide procurement staff and suppliers with training on responsible AI use in procurement, covering how AI is currently used, how bias is recognized and mitigated, how human oversight is maintained, how to escalate concerns or challenges about AI outcomes and measures for protecting supplier data privacy in the context of AI.

Intermediate Progress

Provide targeted training to suppliers on more detailed responsible AI topics, such as how to identify and mitigate bias, how to design and maintain human oversight mechanisms for AI-assisted decisions, how to document AI use in ways that are transparent and auditable and how to manage AI-related risks in their own supply chains.

Advanced Practice

Support strategic suppliers in building their own internal responsible AI governance capability, including train-the-trainer programs that enable them to deliver AI ethics and governance education to their own staff and cascade requirements through their own supply chains.

Data

Initial Action

Work with suppliers to collect baseline information on how suppliers use AI when processing or generating data relevant to the buyer.

Intermediate Progress

Work with suppliers to collect information on their responsible AI governance, including policies and incidents.

Advanced Practice

Work with suppliers to develop reporting on responsible AI outcomes and AI‑incidents.

Advisory

Initial Action

Provide suppliers with clear guidance on how AI is used in the buyer's procurement processes, what role it plays in decisions that affect them and how they can request explanation or raise concerns about AI-influenced outcomes without fear of commercial retaliation.

Intermediate Progress

Provide advisory support to suppliers on strengthening their responsible AI governance, including guidance on conducting bias audits, designing escalation and challenge mechanisms for AI-influenced decision, involving human review and communicating transparently with their own stakeholders about how AI is used in their operations.

Advanced Practice

Support strategic suppliers in embedding responsible AI governance into their own procurement and operational frameworks, including supplier screening criteria for AI use, contract provisions for AI transparency and incident disclosure and monitoring systems for AI-related risks across their own supply base.

Financial Support

Initial Action

Providing guidance resources / template policies or sponsor training to reduce the cost of implementing responsible AI practices for suppliers.

Intermediate Progress

Recognize supplier investment in responsible AI governance in commercial terms, such as through preferential payment terms or reduced audit burdens for suppliers that demonstrate compliance, reducing the indirect financial cost of meeting the buyer's AI governance expectations.

Advanced Practice

Link commercial incentives such as contract extensions, volume increases and pricing adjustments to verified progress on responsible AI governance, including completion of bias audits, implementation of human oversight frameworks and demonstrated cascading of AI accountability requirements through the supply chain.

Audits

Initial Action

Conduct internal audits of the buyer's own AI use in procurement, reviewing whether human oversight is being applied consistently, whether AI tools are producing fair and unbiased outcomes across supplier groups, whether suppliers are being informed of AI use as required, whether all AI tools in use are in the approved inventory and whether escalation mechanisms for AI-related concerns are functioning effectively.

Intermediate Progress

Ask suppliers to complete a self-assessment of their responsible AI governance practices, including:

  • AI use policies
  • human oversight mechanisms and human-in-the-loop requirements
  • bias testing practices
  • algorithm transparency
  • incident response procedures and incident history
  • disclosure of AI use to their own supply chain partners
Advanced Practice

Commission independent assessments of responsible AI governance practices across strategic suppliers' own operations and supply chains relevant to the buyer's risk, reviewing policy implementation, bias audit outcomes, incident history and remediation and the effectiveness of human oversight mechanisms.

Project Partnerships & Innovation

Initial Action

Engage with procurement networks and standards bodies to share good practice on responsible AI use in procurement and align on common disclosure and transparency standards.

Intermediate Progress

Join or establish a cross-industry working group with procurement peers to share intelligence on emerging supplier AI risks, including incidents, governance failures and new AI use cases that create exposure for buyers across shared supply chains.

Advanced Practice

Partner with industry peers, civil society organizations and relevant partners to develop systemic responses to the governance challenges posed by AI in supply chains, such as shared frameworks for AI transparency and accountability and joint research into emerging AI risks in supply chain contexts.

Continuous Improvement

Metrics

Initial Action

Track the following:

  • percentage of procurement staff trained on responsible AI use
  • percentage of AI-assisted procurement decisions documented with human review records
  • number of supplier complaints or concerns about AI-influenced procurement outcomes
Intermediate Progress

Track the following:

  • percentage of AI-assisted procurement processes audited for bias and fairness
  • number of bias audit findings identified and remediated
  • percentage of suppliers with documented responsible AI governance policies
  • aI-related incident reports received from suppliers and resolution rates
Advanced Practice

Require strategic suppliers to report on responsible AI governance performance, including the number of AI-related incidents reported and resolved, progress against agreed governance improvement milestones and the percentage of their own key subcontractors that have disclosed their AI use and meet minimum responsible AI standards.

Scorecards

Initial Action

Include responsible AI practices as a component in supplier scorecards, covering existence of a responsible AI use policy, whether the supplier has disclosed their AI use to the buyer and whether they have a human oversight mechanism for AI-assisted decisions that affect the buyer.

Intermediate Progress

Rate suppliers on the quality and maturity of their responsible AI governance practices, scoring them on policy documentation, bias audit completion, human oversight mechanisms, incident response capability and improvement against previous scorecard results.

Advanced Practice

Weigh responsible AI governance performance in strategic supplier scorecards, with top performers receiving preferential weighting in tenders and bottom performers required to submit improvement plans with defined timelines and escalation consequences.

Impact Verification

Initial Action

Gather feedback from suppliers on whether the buyer's use of AI in procurement is perceived as transparent, fair and explainable, and whether suppliers feel they have effective recourse when they disagree with an AI-influenced outcome.

Intermediate Progress

Assess whether supplier engagement and capacity building on responsible AI governance is meaningfully reducing AI-related risks and improving governance maturity among suppliers.

Advanced Practice

Measure whether the buyer's responsible AI programs are creating demonstrable improvements in supply chain resilience over time, such as whether the number of AI-related incidents and bias findings is declining over time.

Sources consulted

Data Protection & Privacy SME Enablement