AI Vehicle Damage Appraisal: Insurers Are Catching Up

Tesco has cut fleet damage costs by 18% using an AI inspection gantry. DHL and Culina have now joined the pilot. The interesting part isn't the fleet side. It's what happens to the claim once that evidence lands on a desk that's still working from spreadsheets and email attachments.

What's actually happened

Sopp+Sopp's Fleet Scout, an AI-powered vehicle inspection system, has moved from a single fleet pilot to multiple. Tesco has been running it since a pilot began around 18 months ago at its Peterborough distribution centre, and has now reported an 18% year-on-year reduction in damage-related costs. DHL and Culina have since started their own pilots on the same platform. Tesco is reportedly considering extending it beyond HGVs to vans and potentially company cars.

The mechanics are simple. Vehicles drive through a scanning gantry, the system automatically detects and documents damage, and a team of engineers grades the images to keep training the model. What used to be a walk-round with a phone camera is becoming a standardised, timestamped, machine-read record, generated the moment the vehicle passes through.

Why this is a claims story, not just a fleet story

Every credit hire and motor claims operation depends on the quality of the damage evidence it receives at the start of a claim. Right now, that evidence usually arrives as photos, a handwritten note, or a description in an email. It's fine, most of the time, until it isn't, and the moments it isn't fine are exactly the moments that end up disputed.

An automated, documented inspection changes what good evidence looks like on the other side of the desk. If a fleet operator can produce a timestamped, AI-graded damage record the moment an incident happens, the claims operation that receives it is being handed a stronger starting point than it's used to working with, and the claims operation that's still asking for photos over email starts to look like the weak link in the chain rather than the fleet.

The uncomfortable version of this argument

This isn't really about Fleet Scout specifically, and it isn't really about Tesco, DHL or Culina. It's about the fact that automated inspection has now gone from a single pilot to three large operators inside about 18 months. That's the usual shape of a technology moving from novelty to expectation, and it usually moves faster on the second lap than the first.

The operators that get ahead of this are the ones already thinking about what their own evidence-capture process looks like at the point of loss, not the point of claim. The ones that don't will eventually be explaining to an insurer, a broker, or their own board why their evidence base looks thin next to a fleet operator's, which is the same argument we made about evidence capture ahead of GTA ADR, arriving from a different direction.

What to actually do with this

You don't need to build an AI inspection gantry to take the underlying lesson. The point isn't the specific technology. It's that the gap between evidence captured properly at the point of loss and evidence reconstructed after the fact is where claims get slower, more disputed, and more expensive to run. Wherever your own process still depends on someone remembering to take the right photo at the right angle, that's the gap worth closing first, before someone else's automated system makes the comparison for you.

That's also the same problem we've written about in a different shape, when one supplier owns recovery, repair and mobility: the operators who can prove their evidence and their data are the ones who keep winning the work, regardless of which side of the desk they sit on.