A global vehicle-parts distributor · Automotive aftermarket · B2B + B2C
Rated best in the industry for finding rare parts — and still fighting a trust deficit that ate the advantage alive.
Client identities withheld by design. Engagements delivered as lead across previous consulting roles; figures are each engagement's findings and committed design targets.
The situation
A recycled-parts giant with an availability network competitors couldn't touch — customers consistently rated it first for finding rare parts. And they still didn't trust it. Quality disputes clustered around part grading, warranties were opaque, and every wrong-fit part cost a repair shop two to three hours of bay downtime. On the yards themselves, the operating reality broke every standard fix: noise above 80 decibels defeated voice input roughly seven times out of ten, grease and wear made a third of barcodes unreadable, and connectivity dead zones stalled cloud-dependent scans.
What the diagnosis found
Four personas, one wound
Voice-of-customer analysis across the mechanic, the bulk reseller, the DIY buyer, and the warehouse staffer traced every persona's sharpest pain to the same root: nobody could prove a part's condition before it mattered.
The sentiment scorecard
Weighted sentiment synthesis showed availability strongly positive while quality-fit accuracy, data reliability, and returns transparency ran negative — trust was the one theme losing on every front.
The benchmark told on us
The category's gold-standard competitor won on a triad of fitment guarantees, delivery visibility, and ease. The gap wasn't cosmetic — it was structural, and professional buyers were pricing it in.
Labels lied
Yard conditions meant any solution built on listening or reading — voice input, barcodes, printed labels — would fail exactly where it was needed most. The diagnosis ruled out the default answers before designing the real one.
The redesign
Geometry became the language of truth
An AR-first intake engine: capture the part's 3D geometry, match the point-cloud against a master CAD library, and identify by physical shape alone — offline-capable, and indifferent to grease, noise, and missing labels.
Orders validated to the build sheet
AI VIN-decoding checks twenty-plus technical data points before an order exists, ending the culture of "close enough" guesses that produced wrong-part returns.
The digital handshake
A 360-degree visual QA record at the warehouse, sent to the customer for condition sign-off before the truck is loaded. No more surprise rust, no more he-said-she-said disputes on arrival.
A warranty that carries its own truth
Condition captured and stored at the moment of sale, claims opened by scanning the invoice QR, validation instant, resolution self-service. The proof burden disappears because the proof already exists.
What the blueprint committed to
33%
throughput improvement designed into intake
30s
target intake time, against a 45-second legacy process
97%
part-identification accuracy by geometry
When noise beats voice and grease beats barcodes, geometry is the only witness left. Every part should carry its own truth.
The first principles at work
Recognize your own journey in this?
Tell Exie what feels similar — the same diagnostic method maps your version of it. Free, and you keep the map.
Next case · A nationwide landscape-supply networkFrom supplier to partner.