DecorRight, painting and decorating, Perth
Purchase orders from builders and insurers become jobs in one click, not by retyping.
An operations platform for a painting company. AI reads purchase order PDFs, the office confirms, and crews run the job from their phones.
- AI PO intake
- Job pipeline
- Quotes and invoices
- Crew mobile app
467
real purchase orders used to build and test the reader
330
production rates derived from past jobs
Before
- The team ran on an old, clunky ServiceM8 workflow.
- Purchase order PDFs from insurers and builders were keyed in by hand.
- There was no live tracking of production rates or job margin.
After
- AI reads each purchase order and shows its confidence for every field.
- The office checks the draft and clicks Confirm Job.
- Crews clock on, run timers and log notes and photos from site.
- Estimated and actual costs roll up into live job margin.
What we built
Purchase order intake
A three-pane review screen shows the source PDF beside the parsed fields, matched to the rate library.
Pipeline, schedule and map
Jobs move from quote to paid on one board, with crews scheduled by day and location.
Crew app
Today's jobs, shifts, task timers and routes on the phone, with GPS clock-in.
Quotes and invoices
Quotes, proposals and invoices shared by link, with invoicing linked to Xero.
Built with Next.js, Convex, WorkOS, OpenAI, Xero.
2 AI agents in this build
PO reader
Reads purchase order PDFs, double-checks suspicious line items and matches them to the company's rates.
Receipt reader
Pulls supplier, amount and job from receipts so expenses land on the right job.
AI suggests and checks. A person on the team makes the call.
Find out what your business could run on.
Tell us what slows your team down. We will show you what we would build and what it would take.
- 30 minutes on video, at a time that suits you
- A straight answer on whether software will help
- A rough idea of cost and timeline
- No obligation, and no hard sell