Guide · Insight
Preparing for AI on the buying side
Published 2 September 2026
Machine written bids are now mainstream. That changes three things on your desk: volume, similarity and verification. This page is about all three.
2026 procurement trends now firmly note that buyers should assume some bid content is machine assisted. The advice is to redesign questions to test ‘project-specific understanding, delivery capability and evidence rather than generic narrative’.
Procurement lawyers are flagging similar challenges. The same models, on the same ITT, means responses ‘could become very similar or even identical’. The demand is ever increasing for skilled evaluators and nuanced scoring.
PPN 017 is the Government’s posture. AI use by suppliers is not prohibited, transparency plus due diligence is how it’s handled. It also wants buyers to plan for ‘an increase in volumes of responses’, as AI writing makes it easier than ever to submit.
The Government’s Artificial Intelligence Playbook sets 10 principles and practical checklists. I won’t note them down, but here is the link.
Below are some practical tips from what I have seen in the market.
Preparing your buying function
Adopt PPN 017, even if you aren’t bound by it (yet). The template questions and due diligence playbook are written and free, ready for your next ITT.
Plan for more bids as AI will collapse the cost of bidding, allow more time for due diligence and handling higher response volumes. Anticipate more clarification questions and build all of this into procurement timetables.
Look at how you can redesign questions to be ‘un-promptable’. Generic questions will give rise to generic AI answers. Anchor your questions in specificities, such as network detail, depot geography, performance metrics and then force bidders to show their own evidence against it.
Keep your evaluators human and acutely aware of sameness. If three bids read alike, that isn’t coincidence, it’s now an expected failure mode. Reward specificity, verifiable evidence and human flair. Don’t fight AI bidding with AI marking, the legal commentary warns against it, with black box marking inviting challenge, so keep evaluation rationale articulable and defensible. Machine assistance for admin is reasonable, but keep the audit trail.
Lead from the front and set the AI rules first, before suppliers get to take the reins. Disclosure can cut both ways, any buying function with a written position on its own AI use holds a stronger line when governing everyone else.
What to ask suppliers
Ask the straightforward question… ‘Have you used AI or machine learning tools, including large language models, to assist in any part of your tender submission?’ It’s a template within PPN 017.
But don’t score the disclosure. Guidance states AI use is not prohibited and commentary states that disclosure answers should not be marked. Asking the question above is giving you the ability to look closer, not penalise honesty.
Then, the follow up can set you apart. You want to be able to separate governed AI use from grey AI use. Ask for an AI log. Which tools, which models, which sections, which figures, who signed off. Suppliers using AI the governed way will be able to answer, suppliers existing in the grey area will not.
AI and figures is an interesting area. Hallucination with AI use is real and dangerous and it can impact you significantly. Ask bidders to source trace all figures, back to something a human can point at and guarantee. For any claims or statements, ask for the citation or source point. AI is great at making things look perfectly compliant, when behind the scenes, it could be invented.
The tools at your disposal can help. Clarification questions, supplier presentations, site visits, references and case studies. A 30 minute presentation with a supplier will beat the machine, people in the room will reveal more than another bidding round.
AI in bidding is not a threat to good procurement, ungoverned AI is. Authorities can lead the line on AI usage and ensure they stay one step ahead.
Honest advice and guidance is always available if you need it.
For how we work with public authorities, see our work with authorities.
For the supplier’s side of the same shift, read tuning AI to win transport bids.
Sources
- RLB Procurement Trends Report 2026, AI in Tender Responses. rlb.com view source ↗
- Browne Jacobson, AI in public sector procurement (and Local Government Lawyer feature by Anja Beriro). brownejacobson.com view source ↗
- PPN 017, Improving Transparency of AI Use in Procurement (Procurement Pathway). procurementpathway.civilservice.gov.uk view source ↗
- Artificial Intelligence Playbook for the UK Government, GDS. gov.uk view source ↗