Track what changed in a supplier market scan
A wider supplier search can find viable offers that already existed. Keep the membership and criteria of each scan so procurement can separate better discovery from real changes in supply.
Bring sources together, trace their connections and develop a working analysis.
The UK government’s 2025 AI sector study found a much larger population of suppliers than its previous study. It also used improved discovery and classification methods. The procurement lead who wants to know whether a market now offers more viable choices needs to distinguish those changes. A bigger supplier list can be valuable even when part of the increase comes from looking more effectively.
The study, published on 29 October 2026, maps firms with evidence of commercial provision of AI products and services and a UK registration. It measures the supplier side of the market. The buyer’s question is narrower. For a defined work package, can the organisation now invite credible bids from suppliers it could not previously reach?
Keep the count and its membership
The study identified 10,488 in-scope firms for 2025, compared with 5,862 for 2024. The 2025 population includes 4,580 firms from the earlier study that were still active and in scope and 5,908 brought into the dataset. The report distinguishes market growth from changes in coverage, classification and how companies describe their offers. The additional records should not all be read as newly established businesses. (UK AI sector study, section 2.2 and Appendix 4)
A buyer can apply the same distinction to a market scan. Retain the earlier supplier list and the rules that produced it. When refreshing the search, match legal entities and record changes in the specific offer being considered. A company may appear for the first time because it launched a product, because it was newly discovered or because the work package has changed. Give those cases separate states. Where the evidence is incomplete, retain that uncertainty.
Consider a hypothetical transport operator seeking a service to extract equipment identifiers from maintenance records. The first search examined established document processing suppliers. A refresh also searches specialist industrial software and discovers an existing vendor whose product already handles the relevant record format. That is a useful new candidate for the operator. If the vendor’s offer existed before the first scan, it is evidence that the search improved. A different vendor whose new release adds the required format is evidence of changed supply. Either may deserve an invitation, subject to the same checks.
A new listing can have several causes
Swipe or scroll for the full diagram →
A change in list membership warrants a reason.
- Observed
- In this example, the search finds an existing industrial product and a separate newly released capability.
- Inference
- Discovery coverage improved and one supplier's offer changed.
- Evidence boundary
- Whether either offer meets the buyer's requirements on its equipment records.
A proposed comparison of successive searches for an equipment record service.
An existing offer discovered by a wider search and a new product release are both candidates for procurement. The first changes search coverage and the second changes available supply. Both still require capability evidence.
Separate search coverage from buying readiness
For the proposed procurement, define what a viable offer must demonstrate. The transport operator might require extraction of identifiers from poor scans, a passage or image region for each result and a route for unreadable records to reach a reviewer. A broad product description that mentions AI helps discover candidates. A dated product document or agreed demonstration on representative records can establish which function exists and whether it meets the operational requirement.
Keep product identity alongside legal identity. Two catalogue entries may describe the same product through different resellers. A supplier with several relevant products may warrant several comparisons but still represents one legal counterparty. The buying team should choose whether it is counting companies, products or usable routes to obtain the service and apply that unit consistently.
The sector study’s methodology uses commercial and public sources, web evidence and manual review. It describes its assessment as a point-in-time view with final checks in June 2026. That makes the study a starting point for market discovery. The timing and scope of the transport operator’s work package still determine which supplier evidence it needs to verify directly.
Retain the comparison when the brief changes
If the operator broadens the brief to accept externally hosted processing, show what the revised rule adds separately from changes under the original rule. If the team can reconstruct both scans with consistent criteria, it can compare the result for those criteria over time. If earlier evidence is too thin to do so, mark the comparison as incomplete. Keep candidates that failed an earlier test visible with the reason and evidence date, so a genuine new release can trigger reconsideration.
This is an upstream decision from reassessment during an AI procurement. Before assessing how an appointed supplier responds to change, the buyer needs to decide which offers are worth putting through the same selection process. A maintained market record also helps the team explain why a wider search produced few extra candidates, if that is the result.
The procurement lead should use the updated scan to select the offers that warrant the same evidence request. If discovery widened but the group of viable offers stayed small, use the capability gaps to guide further market engagement. If the qualified group grew, invite the added candidates on the same terms and preserve the evidence that changed the decision.