How we built FixPilot, an AI buying co-pilot for auto dealers

An AI co-pilot that estimates a car's retail value and recon cost from the auction listing and its photos, then gives dealers a max bid.

Book a working session

Blueprint drawing of a pickup truck beside a fanned stack of photo prints, a print under a magnifying loupe, routing lines running into a blank price tag

The challenge.

Dealer buyers work through auction runs by hand, scrolling listing after listing and opening every set of photos. To bid well on a car they need to know what it would sell for at retail, what it is going for at auction, and what reconditioning will cost once it lands, and that last figure hides in the photos and the seller's notes. When the estimate is wrong, the dealer buys a car they regret.

What we built.

Pulled photos, VIN and seller notes from a pasted search or listing
Decoded each VIN for trim and factory option packages
Read every photo with a frontier multimodal model for damage and wear
Priced each finding as reconditioning at the store's own rates
Set retail value from sold listings near the lot and wholesale auction data
Computed a max bid per car from retail value, recon, transport and margin

Results of this engagement.

A whole run is graded at once instead of by hand
Retail value and the auction price sit side by side for every car
Recon cost comes from the photos and the listing, not a guess
Every car carries a max bid at the store's target margin
Each rooftop gets its own max bid on the same listing

Where to go next

Bring one broken process.

No discovery marathon. Bring one process that costs your team hours and we will map what an integration would look like in a free 30-minute working session.

Book a working session