Skip to content

ConceptAI app

Walkround: an AI walk-round check app for van and car hire

Staff walk round the van with their phone. The AI marks new damage against the check-out photos and prices it. The customer signs on screen and gets the report.

Try it on your phone

The brief

For: A van and car hire firm with two depots and a damage bill it can't recover

A van and car hire firm does its walk-round checks on a paper sheet with a van outline, and settles damage arguments from blurry photos on staff phones. It wants a guided check on the phone, new damage spotted against the last check, a signature in the yard and a report the customer can't dispute later.

“Walk round. Sign. Done.”

Walkround concept: the walk-round camera and an AI damage result on two phones

Every main screen of Walkround.

Scroll sideways. Each screen is part of the fixed scope on this page.

How we'd build it

The scope we'd quote for Walkround today, as a fixed price paid in stages.

What we'd deliver
Staff app · Customer's phone · Fleet board
Platforms
iOS · Android · Web
Size
Starter
Typical time
About 4–6 weeks

See pricingMobile app development

Walkround review and before-and-after screens, with new damage in amber and old damage in grey
Walkround fleet board in a browser, with the customer signing on their phone

How it works

  1. Walk round the vehicle. The app asks for six photos: front, nearside, rear, offside, roof and interior. An outline shows where to stand, and the photo takes itself when the phone is held still.
  2. The AI checks each photo. Marks are pinned on the photo with a confidence score. Anything that was on the check-out photo stays grey. Anything new pulses amber, with a repair estimate.
  3. Staff confirm or dismiss. A before-and-after slider compares each mark with the check-out photo. Low-confidence marks are flagged as possible dirt, so a sleeve settles it.
  4. The customer signs. On the staff phone in the yard, or from a text link on their own phone. They see the photos and the estimate before they sign.
  5. The report goes out. A printable report with every photo, the marks and the signature lands on the customer's phone and the fleet board straight away.

The hard parts

  • Training the damage model. Off-the-shelf models find dents on clean, well-lit cars. Hire vans are dirty and photographed in the rain. The custom AI work is fine-tuning on the firm's own past checks, and setting the confidence level below which a mark is flagged as "check by eye".
  • Comparing like with like. The outline guides make each photo of a side line up with the last one, so the comparison is fair. Without them, angle and distance changes look like damage.
  • Camera permission and bad light. The camera is asked for at the first check, not at sign-up. The app warns when it's too dark to see damage and falls back to choosing a photo from the phone.
  • Signal in the yard. Photos are queued on the phone and upload when there's signal, so a check never stalls at the back of the yard.
  • Evidence that holds up. Every photo is stored with its time and the check it belongs to, and the report can't be edited after it's signed.

What waits for version two

  • Video walk-rounds, with frames picked out automatically.
  • Linking the estimate to a body shop's booking system.
  • Charging the deposit from the app. In version one the office charges it from the fleet board.

Week by week

WeekOn your phone
0App Preview: the walk-round flow, clickable, plus a fixed quote
1Accounts opened in your name; first test build
2Every screen designed and signed off; your past check photos gathered for the AI
3The guided walk-round and signature working end to end
4Damage detection on your own vans, with the confidence levels tuned
5Submitted to Apple and Google from your accounts
6Live in both stores

The design system

Every concept gets its own brand, so it looks like a product, not a template.

  • Hi-vis amber#FFB21A
  • Graphite#121417
  • Gunmetal#1C1F24
  • Steel#8E97A3
  • Chalk#F2F1EC
Display
Archivo Expanded
Interface
Archivo

Questions

Is Walkround a real company?

No. Walkround is a concept app: a product we designed end to end to show how we'd build an AI vehicle check app for a hire firm. The scope, weeks and price are what we'd quote for it today.

Does the AI decide what the customer pays?

No. The AI suggests marks, with a confidence score, and prices them from the depot's own repair list. Staff confirm or dismiss every mark before the customer sees it, and the repair invoice sets the final figure.

How does it tell new damage from old?

Each return is compared with the photos from that hire's check-out. Marks already on record are shown in grey and never charged twice. Only what's new is shown in amber.

Why Starter and not Growth?

There's one kind of app user, the staff member doing the check. The customer signs on the staff phone or opens a link on their own phone, so they never install anything. The fleet board is the Starter admin dashboard, and the damage detection is the custom AI on top.

More work

All our work
  • ConceptStaffing marketplace

    Shiftbay: Swipe. Check in. Paid Friday.

    Workers swipe to claim shifts near them, check in by location and get paid every Friday. Businesses post shifts and approve timesheets. The first version ran on Bubble; this one is code, with the data moved over.

    Worker app + Business app + Ops dashboard + Data move from Bubble · about 8–12 weeks

    Try it on your phone

Have an app in mind?

Tell us about it on a free 20-minute call. If we're not the right team for it, we'll say so.

Book a 20-minute callSend us a message

A reply within 1 working day. Your App Preview can start within 2.