Share with your COO
Construction companies chasing AI-powered scheduling and autonomous plant are fixing the wrong problem, according to procurement data from YardLink’s Neeral Shah, drawn across more than 19,000 UK projects. For a mid-sized contractor turning over £12m, supply chain failures, failed deliveries, unvetted suppliers, and non-compliant waste documentation, can drain up to 8% of annual revenue before any AI system touches the project. A single failed delivery costs £750 to £800 in labour and programme impact. Procurement admin alone burns £25,000 to £35,000 a year in recoverable capacity. Systematic supplier vetting, by contrast, cuts site downtime by roughly 65%.
What this means for your business
The companies most exposed here are portfolio operators running multiple live sites simultaneously, because the losses compound across sites, not just within them. A single delayed delivery at £800 is a nuisance. The same failure pattern across five sites, multiplied over a project lifecycle, reaches £75,000 in lost productivity before anyone has reviewed a single AI vendor proposal. If your procurement discipline varies by site, by whoever was available that week to vet a supplier, you’re not running a portfolio, you’re running five separate risk exposures with a shared P&L.
The structural problem Shah identifies is real and often invisible in standard reporting. Margin erosion from procurement failure doesn’t show up as a clean budget variance. It accumulates in invoice query volumes, in downtime that gets coded as weather or labour, and in compliance penalties that arrive months after the offending supplier was paid and moved on. Around a third of waste carrier documentation in UK construction supply chains is non-compliant or missing critical information, and the legal duty of care stays with the contractor, not the lorry driver who left site. That’s a liability sitting in your supplier file, not in your risk register.
The argument that AI automation layered on broken procurement just accelerates the losses is correct, and the implication cuts across industries well beyond construction. Any operation where supplier reliability is managed informally, by relationships and habit rather than by verified credentials and systematic checks, faces the same dynamic. The falsification condition worth tracking is this: if structured supplier vetting and automated compliance checks genuinely push invoice query rates below 1.5% as Shah claims, that’s a measurable internal benchmark any COO can run against their own current rate to judge whether this is their problem or someone else’s.
Based on reporting from While you’re chasing AI, procurement is draining your margin, originally published 2026-09-04 09:13:00.
