[ Case study ]
Pre-delivery and warranty inspections lived on paper forms that were photographed, emailed, and retyped — warranty claims waited on office transcription, and disputed inspections had no authoritative record.
CLIENT a four-branch agricultural equipment dealer — FOCUS Design for the shared device
Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.
Four branches of an agricultural equipment dealer spread across two hundred kilometres, selling and servicing tractors, harvesters, and implements. Every machine delivered or claimed under warranty generates an inspection, and inspections were paper: a multi-page checklist completed at the machine, photographed, and emailed to the office, where a service administrator retyped the findings into the dealer management system. Technicians move between branches in peak season, and warranty processing is the difference between being paid by the manufacturer and absorbing the cost.
Pre-delivery and warranty inspections lived on paper forms that were photographed, emailed, and retyped — warranty claims waited on office transcription, and disputed inspections had no authoritative record.
We proposed an inspection app for the branch iPads that models the inspection as structured data — checks, measurements, photos, and signatures as records, not fields on paper. A fast user-switch pattern keeps every inspection attributed through shift handovers, everything queues locally with visible sync state for the metal sheds, and a scheduled export produces exactly the import format the on-premise DMS expects each night. Disputes become reconstructable, and transcription disappears instead of moving to a new screen.
Just as important is what we ruled out, and why:
A fast user-switch pattern with a branch-selected profile keeps each inspection attributed, and the app queues everything locally with a visible sync state.
Checks, photos, and measurements are structured records, so the DMS import is deterministic and a disputed claim can be reconstructed exactly.
A scheduled export produces the DMS's expected import format, making the on-premise constraint a timing detail rather than a blocker.
Delivered by the systems pod — engineer + mobile engineer over 9 weeks, with working increments reviewed with the client every week.
Obstacle
Within the first fortnight of the pilot, full-resolution machine photos filled the shared iPads — technicians photograph serial plates and hydraulic hoses, and nobody was deleting anything.
Handled: We shipped an emergency release with HEIF capture and automatic purge after successful sync, then set a storage alert per device so the next fill would warn before it happened.
Obstacle
The DMS vendor's import spec turned out to be versioned — the on-premise install's schema revision was undocumented, and records written against the wrong revision were rejected silently at the nightly import; the office learned of it only when warranty claims went missing.
Handled: We negotiated a versioned schema agreement with the vendor, so each export declares the schema revision it targets, and added a pre-import validator plus a morning exception report — the service administrator clears flagged records before the warranty run instead of discovering gaps in it.
The headline: all four branches on the app, with warranty claims supported by machine-readable inspection records — Paper forms → structured inspections, read from Warranty-claim processing log. A second check: inspection-to-claim-submission time at 3 days → same day.
The warranty administrator's job stopped being transcription and became adjudication: she reviews machine-readable inspections with photos attached and prepares claims the same day they are inspected. Technicians stopped defending paper against memory — the record is the record, and manufacturer disputes end with a shared inspection file instead of a standoff. Branch managers can see inspection volume by technician, and checklist updates propagate to all four branches the day a model-year change lands.
The result was read from Warranty-claim processing log against the pre-engagement baseline over the stated window, with a guardrail check on inspection-to-claim-submission time. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.
What we would do differently
We would have specified photo compression before the pilot — full-resolution machine photos filled the shared iPads within the first fortnight and forced an emergency release.
[ Related service ]
[ Related builds ]
62% 89%Share of digital tickets validated offline-capable at first launch of the trial season, read from the operator's validation logs
3 tools 1 appMembers manage access, bookings, and billing queries in one app across all five cities
[ Next step ]
Next case study