Independent product concept · TEG / Trustd research · 2026
An AI concept for reducing compliance friction in transport marketplaces.
Concept prototype — not connected to TEG systems or real customer data.
01 — The problem
Works out the requirement, fixes the document, submits again.
Reviews the same file twice, and the queue behind it waits.
A slower path to trading, and an account that can't earn yet.
Potential impacts, not measured figures — each would need validating against TEG's internal data.
02 — The insight
The goal isn't to automate compliance decisions. It's to stop predictable problems before they turn into rework.
03 — The product
Extracted from your document
The vehicle registration on this document doesn't match the vehicle on your profile.
Your policy expiry date appears to be 14 days away.
Catch it at the upload, explain it in one line, and the customer is right on the first attempt.
04 — Try it
Working prototype, sample data. Drag time forward and let a document lapse, or renew one early — one renewal has a deliberate mistake for the pre-check to catch.
05 — Why now
transport and logistics businesses already run their identity and compliance through Trustd.
Reported by Trustd, April 2025platform in transport and logistics certified as an Identity Service Provider under the UK digital identity framework.
Reported by Trustd, April 2025When a profile becomes a certified identity, the document that fails the check stops being admin. It's the thing standing between an operator and the work.
06 — The number
Primary metric
Baseline first, then a controlled pilot on one document type. Build further only if the number moves.
07 — The bets
Most failed submissions contain predictable errors a machine could catch before review.
Sample recent rejections and classify why each failed. Kills it: fewer than half are machine-detectable.
Customers act on a warning at upload rather than ignoring it.
Prototype with a small group of operators. Kills it: most click straight past to “send anyway”.
Pre-validation cuts repeat submissions without letting anything through it shouldn't.
Controlled pilot on one document type. Kills it: resubmissions fall but false clears rise.
Each bet has a number attached and a result that would end it. I'd rather find out cheaply than build on an assumption.
08 — Why me
I built it because during my research into TEG, I found myself asking: if this is the problem, what could we build to make it better?
I don't have TEG's internal data, and I wouldn't pretend to know the business better than the people running it. But I know how I think.
That is the kind of product thinker I am.
Sources