Rinnai says Oracle's AI now reads its emailed orders. Manual order processing fell 20%.
Rinnai Australia makes hot water, heating and air conditioning systems. It handles more than 120,000 customer orders a year, and most arrive as emails or scanned documents in each customer's own layout. Staff keyed many of them in by hand, which Oracle's customer story says risked errors and delayed deliveries.
Rinnai now runs those documents through Oracle's cloud AI service Document Understanding. It pulls tables, bar codes, signatures and order lines out of each one and, Oracle says, creates a standard order in JD Edwards, the business system where Rinnai books its orders. Rinnai taught the AI which fields to pick up.
In Oracle's film, Rinnai's general manager of finance and ICT (IT and telecoms) puts the cut in daily order processing work at "about 20%". He also says the system "has eliminated the risk of errors". Oracle's written story says the change "helped improve data accuracy".
Neither source says how Rinnai checks an extracted order, its customer code, item numbers and prices, before it enters that system.
What checks would you require before an order read by AI enters your systems?
🎥 Video: Oracle, official YouTube channel, 16 Oct 2026 · narration: AI voice
What the captions say
- Rinnai Australia makes hot water and heating systems.
- It gets over 120,000 customer orders a year.
- Most arrive as emails or scanned documents.
- Staff often typed them into the system by hand.
- Now Oracle's AI reads them into its JD Edwards ERP.
- Rinnai says manual order work fell about 20%.
- Both figures come from the companies.
