Why Subcontractor Vetting Fails on UK Projects
Subcontractor failure is one of the most expensive risks in UK construction. A mechanical and electrical subcontractor collapsing mid-programme on a £4.2m commercial fit-out in Manchester doesn't just create a procurement headache — it triggers delay claims, disrupts the JCT programme, and exposes the principal contractor to liquidated damages that can run to £5,000 per week or more. Yet despite the stakes, most GCs still vet subcontractors using the same manual checklist approach they used a decade ago: request a few documents, skim the accounts, tick the boxes, and move on.
The problem isn't laziness — it's volume. On a live project pipeline, a procurement manager might be evaluating four or five subcontractors simultaneously across different trades. Properly reviewing a subcontractor's financial accounts, insurance certificates, RAMS templates, Building Safety Act 2022 compliance documentation, and previous project references takes hours per submission. Corners get cut. Red flags get missed. And when something goes wrong six months into the job, the warning signs were there in the documents all along.
AI-powered document analysis is changing that equation — not by replacing professional judgement, but by doing the heavy lifting on document review so procurement teams can focus on the decisions that actually require human expertise.
What Documents Actually Matter — and What They're Hiding
Effective subcontractor vetting on UK projects requires scrutinising a specific set of documents, each carrying its own risk signals. The challenge is knowing what to look for inside each one.
Companies House filings are the starting point. Abbreviated accounts can mask deteriorating current ratios, overdrawn director's loan accounts, or net liabilities buried in the notes. A groundworks subcontractor quoting £380,000 on a civils package might look solvent on the surface but be carrying £220,000 in creditor pressure that makes them a retention risk before practical completion.
Insurance certificates deserve more scrutiny than most procurement teams give them. Check the retroactive date on professional indemnity cover, confirm that the policy covers the specific scope being tendered, and verify that the insurer is FCA-authorised. A certificate that expired three weeks ago — or covers "general building works" when the scope is structural steelwork — is worse than useless.
Under the Building Safety Act 2022, subcontractors working on higher-risk buildings must demonstrate competence in ways that go beyond the old CSCS card check. Gateway 2 and Gateway 3 documentation requirements mean that procurement teams now need to assess whether a subcontractor's quality management processes, incident reporting, and golden thread obligations are genuinely in place — not just asserted in a pre-qualification questionnaire.
The PQQ Confidence Trap
Most subcontractors answer pre-qualification questionnaires optimistically. They tick "yes" to ISO 9001 certification that lapsed eighteen months ago, claim experience on comparable projects they only contributed to peripherally, and submit insurance certificates without checking whether the indemnity limits actually meet your contract requirements. Manual review rarely catches these discrepancies — AI document comparison does.
How AI Document Analysis Strengthens the Vetting Process
The practical application of AI in subcontractor vetting isn't about replacing your procurement team's experience — it's about giving them better information, faster. As we covered in our guide to mitigating subcontractor risk, the goal is evidence-based decision-making, not algorithmic selection.
Trueleveler's Bid Scope Compliance engine can be applied directly to subcontractor submissions — upload the tender scope against a subcontractor's returned bid and the engine flags missing items, pricing gaps, and scope exclusions that would otherwise surface as variation claims mid-contract. On a £650,000 M&E package, catching a subcontractor who has excluded builder's work, commissioning, and testing at tender stage saves a dispute that could easily run to £40,000 or more.
The Contract Review engine is equally valuable in the vetting phase. Before you issue a subcontract — whether JCT Design and Build 2016 or NEC4 Engineering and Construction Subcontract — you can run the draft agreement through the engine to identify clause-level risk: back-to-back payment provisions, net contribution clauses, or fitness-for-purpose obligations that your subcontractor may not have the insurance to support.
Catch Scope Gaps Before You Award the Subcontract
Trueleveler's Bid Scope Compliance engine checks subcontractor submissions against your specification — flagging missing items and pricing gaps in minutes, not hours.
Reading the Contract Risk Signals in Subcontractor Submissions
Beyond financials and insurance, the documents subcontractors submit at tender stage often contain contractual risk signals that are easy to miss under time pressure. Pay particular attention to qualifications and exclusions buried in covering letters or programme assumptions buried in method statements. These form part of the tender and, if not struck out before award, can be incorporated into the subcontract by conduct.
That qualification is legally significant. The Housing Grants Construction and Regeneration Act 1996 does prohibit pay-when-paid clauses in most circumstances — so the subcontractor is on solid statutory ground there. But the 14-day payment demand conflicts with your JCT subcontract's 21-day payment period and your own cash flow model. If this qualification isn't addressed before award, you have a payment terms dispute waiting to happen. As explored in our post on contract clauses every GC should red-line, these embedded qualifications are among the most common sources of post-award disputes on UK projects.
Adjudication rights under the 1996 Act mean that a subcontractor who feels underpaid can trigger a referral within 28 days of a notice of dispute — and you'll be defending that referral while the project is still live. Catching payment term conflicts at vetting stage costs nothing. Defending an adjudication costs £15,000 to £40,000 in legal and expert fees, minimum.
Best Practice: Document Comparison Before Award
Run a structured comparison between your issued subcontract terms and the subcontractor's returned submission — including any covering letter qualifications — before signing. Trueleveler's Document Compare engine catches price drift and unauthorised changes between documents, making this a five-minute task rather than a two-hour legal review.
Building a Repeatable Vetting Framework
The goal isn't a one-off document review — it's a consistent, defensible vetting process that holds up to scrutiny from your client, your insurer, and, if it comes to it, an adjudicator or court. That means the same document checklist applied to every subcontractor in a trade package, with AI-assisted analysis creating a documented audit trail of what was reviewed and when.
Structure your vetting around four document categories: financial health (Companies House, latest filed accounts, credit report), insurance adequacy (EL, PL, PI — checked against contract requirements, not just existence), competence evidence (Building Safety Act 2022 compliance where applicable, relevant project references, supply chain capacity), and commercial submission (bid against scope, qualifications, programme assumptions). As detailed in our guide to mastering bid leveling, the commercial submission review is where AI delivers the most immediate time saving — comparing multiple subcontractor bids against a consistent scope benchmark in minutes.
Use AI to process and flag. Use your procurement team to decide. That division of labour is what makes AI-assisted vetting genuinely scalable across a multi-project pipeline.
The Bottom Line
Subcontractor vetting on UK projects has never been a paperwork exercise — it's a risk management discipline. The documents tell you what a subcontractor is actually capable of delivering, how financially resilient they are under programme pressure, and whether the commercial terms they've submitted are compatible with your own contractual obligations up the chain. The problem has always been that reading those signals properly takes more time than most procurement programmes allow.
AI document analysis closes that gap. It doesn't make the procurement decision for you — but it ensures that when you make that decision, you're working from complete, analysed information rather than a hurried skim of a submission pack. On a £500,000 subcontract package, the difference between a well-vetted award and a poorly-vetted one can easily be six figures in dispute, delay, and remediation costs. The technology to do this properly is available now, it's fast, and the cost of not using it is measurable.