Skip to content
Five broad weathered ochre, slate, mauve, terracotta and ivory pigment pieces interlock along an off-centre diagonal seam.

Public sector, education and nonprofitsArticle

What changes when public procurement controls AI tool choice?

Connect public AI purchasing routes to the service staff can use. Separate supplier listings, bundled features, local acceptance and usable records.

Jump to a section

Public procurement changes AI tool choice by making the purchasing route, agreed service and evidence for use part of the adoption decision. A public-service team needs to know what its organisation can buy, what the supplier has actually committed to provide and which staff tasks the responsible owners have accepted. A supplier listing or an available feature cannot answer all three questions.

Start with the service you need and the authority to obtain it. Then connect the purchase record to the instructions employees will follow. This article uses UK government guidance to examine that handoff. The applicable rules depend on the public body, jurisdiction and procurement, so involve your commercial colleagues before choosing a route or changing an existing arrangement.

Establish what the purchasing route gives you

An approved route can make it easier to find relevant suppliers. It still has a defined process and scope. For example, the Government Commercial Agency's Artificial Intelligence Dynamic Purchasing System describes shortlisting suppliers, inviting them to a further competition against a detailed specification and evaluation criteria, then evaluating responses and awarding a contract. Appearing on that system is a step in purchasing, rather than a complete local adoption decision.

Ask your commercial owner to identify the actual agreement and process that apply to the proposed service. If the organisation already has a contract, establish whether the request fits it and how changes are handled. Do not infer permission to add a feature from a supplier's presence on another agreement.

The UK government's 2025 AI playbook recommends engaging commercial colleagues early and starting requirements with the problem, data needs and ongoing support. That changes the adoption lead's first job: describe the work clearly enough for the purchasing team to assess the request.

For a proposed staff assistant, write down:

  • The task and users. What employees need to complete and who receives the result.
  • The information. What would enter the service, where it comes from and which handling questions remain open.
  • The required service. The relevant features, access, support and connections, including what the supplier must demonstrate.
  • The decision requested. A bounded investigation, a purchase, a contract change or permission for a defined use.

Keep general supplier diligence in the AI vendor evaluation process. The additional public purchasing question is how this specific request can be obtained and governed through the organisation's applicable arrangements.

Notice when AI arrives inside an existing service

A team may encounter AI through a product it already uses. A new transcription option in a meeting suite can look like a minor feature, while introducing new questions about information processing, records and access.

Procurement Policy Note 017, published in February 2025, explicitly discusses AI in services that were not originally described as AI purchases. Its video-conferencing example raises questions about generated meeting records and their subsequent use.

Keep the note's scope and purpose clear. It applies to UK central government departments, executive agencies and non-departmental public bodies; other contracting authorities may choose its approach. Commercial colleagues should check which version applies to the procurement.

Its example AI disclosure questions are optional and for information, not scored. They distinguish AI used to prepare a bid from AI used to deliver a service. Neither disclosure is a blanket approval for employees to use the resulting product.

Turn feature availability into a specific request

Suppose Jamie Lee coordinates an estates team's office-refurbishment status meeting at a public body. The team uses an approved video-meeting service. The supplier now advertises AI transcription as included in the licence, and Jamie wants a draft action note that the meeting chair can check and file in the agreed records location.

The existing local record covers video meetings. It does not yet resolve the proposed transcription feature's information handling, access or record export. Jamie should preserve those unanswered questions. Copying a colleague's “we already have it” message into the rollout brief would conceal the distinction the commercial and information owners need to assess.

In a May 2026 Reddit discussion among Canadian public servants, one commenter qualified another person's tool-approval statement by explaining that their own department lacked the relevant enterprise licence. Other replies described different configurations. These anonymous accounts do not establish any department's current permissions or a UK purchasing rule. They illustrate the employee's problem: an approval statement from somewhere else may leave the local question unanswered.

What Jamie can establish about the proposed meeting feature

Existing service

Video meetings accepted

The current local record covers the ordinary meeting service.

Supplier offer

Transcription advertised

Verify the included feature and applicable terms with the commercial owner.

Information handling

Live use still open

Resolve processing, access and records conditions for the proposed meeting task.

Usable record

Export still untested

Request an authorised rehearsal of chair review, correction and handover.

Separate what the current record supports from what the proposed feature still needs. An advertised addition does not complete the local use decision.

For Jamie, “included” is useful pricing information to verify. It does not close the feature request. The next conversation should identify the current contract, the proposed use, the unresolved evidence and the person who can decide each point. Continue the existing note-taking arrangement while those questions are open. A practice test with invented meeting material is a separate request that also needs the appropriate approval.

Keep the purchase record connected to staff use

The adoption lead should be able to explain how the purchasing decision reaches the person running the meeting. Give each record a clear purpose so staff do not have to interpret contract wording during their work.

RecordQuestion it answersWhat staff still need
Purchasing decision and agreement
What service has the organisation obtained, through which arrangement?
The permitted local task and configuration
Information and service acceptance
Which inputs, access and outputs have the responsible owners accepted?
Practical instructions and a route for exceptions
Staff task guidance
How should the employee complete and check this work?
Current support and notice when the accepted scope changes

These records should point to one another. They serve different readers, so repeating the entire contract in a training guide would make updates harder without helping Jamie decide whether to start transcription.

For Jamie's meeting task, the staff guidance should answer:

  • Where does this apply? Name the accepted meeting types and permitted material.
  • Who can see the draft? Identify the agreed access and where questions about it go.
  • Who accepts the record? Explain the chair's checking and correction responsibilities.
  • Where does it go? Name the agreed destination for the checked version.
  • What happens if it fails? Give the ordinary fallback and escalation route.

Have the information and records owners confirm the actual participant-information and retention requirements. Those conditions need to survive the handoff into staff guidance.

When approval is limited, make the limit usable. “Approved for a dummy rehearsal, live meeting information still awaiting review” tells a coordinator which work can proceed. An unexplained green tick beside the product name does not.

Test whether colleagues can complete the record journey

Acceptance needs to cover the work after a draft appears. Can the meeting chair inspect and correct the note? Can the coordinator reach the controls using the input method they need? Can the checked record be retrieved in a usable form by the people who are supposed to receive it?

The government's accessibility testing guidance combines manual and automated checks with assistive-technology testing and research with disabled users. A polished supplier demonstration cannot replace that work. Define the relevant staff journey and involve accessibility specialists and users early enough to change the requirement or choice.

An employee uses a keyboard at a mobile lectern in an empty meeting room, rehearsing how to retrieve a meeting record from a screen facing her.
A rehearsal follows the record beyond its first draft. It can expose a usability question without establishing that the service meets every accessibility requirement.

In the LinkedIn discussion beneath Aditya Katira's October 2026 procurement article, AI consultant and product-studio cofounder Aish Shrivastava recommended trying a sample evidence export against a failed run. He was offering advice from his consultancy and product-development work, rather than reporting a measured public-agency result. It draws a useful distinction for this task: a document describing an export feature and a record the team can actually use are different evidence.

For an authorised rehearsal, ask the relevant reviewers to examine:

  • Access to the task. Can the intended staff members use the relevant controls and information?
  • Review and correction. Can the chair find an incorrect action or attribution and preserve the checked version?
  • Retrieval and handover. Can the coordinator place the accepted record in its agreed destination, with the necessary context?
  • Failure and recovery. Can staff continue through the ordinary process when transcription or export is unavailable?

Record an unresolved barrier with its owner and the affected use. A supplier's proposed fix needs a recheck. Do not average a colleague's inability to complete the task into a favourable overall satisfaction score. For the broader service-value decision, evaluate public outcomes and staff effort together.

Preserve room to change the service later

The purchase also shapes future choices. If action records can only be understood inside the supplier's service, moving away may require more than cancelling licences. Establish which records, context and support the organisation needs to keep, and what the agreement actually provides.

The Digital, Data and Technology Playbook's exit guidance recommends early planning for knowledge transfer, support and transition. Apply that planning to the work people will depend on. Have the commercial, technical and records owners examine the practical handover instead of treating a downloadable file as proof of a usable exit.

The older AI procurement guidelines also connect implementation to training, ongoing support and end-of-life arrangements. An adoption programme therefore needs a way for staff problems and supplier changes to reach the people managing the agreement.

For Jamie's meeting feature, retain the accepted configuration and review date beside the operating guidance. Agree who receives notice of material feature or processing changes, which conditions need reassessment and how ordinary note-taking continues during review. A useful purchasing decision leaves the team able to explain its present choice and reconsider it when the evidence changes.

Questions about public procurement and AI tool choice

Does an approved supplier mean staff can use its AI tool?

A supplier's inclusion on a purchasing arrangement does not by itself establish permission for a particular employee task. Identify the agreement, purchased service and local acceptance decision. For example, the Government Commercial Agency's AI purchasing system describes a further competition and award after supplier shortlisting. Ask the commercial and service owners which arrangement and permitted use actually apply to your team.

Can a bundled AI feature skip the review?

An included feature still needs assessment against the existing agreement and proposed use. A transcription option may change what information is captured, how it is processed and which records people can access. Ask the contract, information and service owners to resolve the affected questions before enabling that use. An addition included in the licence price still leaves those questions open.

Does PPN 017 ban AI-written bids?

No. PPN 017 says suppliers' use of AI during the commercial process is not prohibited and discusses proportionate due diligence. Its example disclosure questions are for information and should not be scored. Commercial colleagues should establish which guidance applies to the procurement and design any further evaluation questions through the proper process.

What should employees do while a feature is awaiting approval?

Continue the accepted way of completing the work and send a specific request through the organisation's normal route. Name the task, feature, proposed information and unanswered question. If a limited rehearsal could resolve uncertainty, request that scope explicitly and use only the material it permits. Another team's access or a supplier demonstration cannot authorise your live use.

How can the team reduce dependence on one supplier?

Plan how the work would continue under another arrangement, then check whether the proposed terms and service support that plan. Examine record formats, required context, access, transition assistance and the people needed for handover. Rehearse an appropriate export and retrieval task before depending on it. Technical portability is part of the decision; commercial rights and an operable transition still need their own evidence.

Updated

aiready

A home for your company’s AI community.

Share what works and help each other put AI into practice.

  • Real use cases

  • Practical guides

  • Company policies

  • Shared experience

Explore aiready
Explore the blog