3Lines: Teaching Regi to Read Handwritten Notes and Scanned Documents

Brief

3Lines Group had already seen what our AI compliance assistant Regi could do for their Statement of Advice reviews, and came back wanting more – because a huge amount of the real-world documentation compliance reviewers deal with arrives as handwritten notes and scanned forms, not clean typed text, so we expanded our relationship into an ongoing embedded engagement with OCR as the first priority, giving Regi the ability to read handwritten and scanned documents and turning what used to be an 8-hour, $1,600 external consulting job into a 10-minute, $5 one.

Energetic, highly motivated and absolutely true to their word – what they commit to, they deliver… It’s rare to find a team that combines innovation with reliability so seamlessly. We’re incredibly pleased with what we’ve achieved together.

EloiseDirector, 3Lines Group

Problem

Compliance reviewers don’t just deal with tidy digital documents. A huge amount of the real evidence trail in financial advice – meeting notes, annotations, scanned forms – exists on paper or as a handwritten scrawl, and none of it could be read by the AI system reviewers already relied on for typed Statements of Advice. That meant a genuine gap: the moment a document wasn’t clean, typed text, a reviewer was back to doing everything by hand, or paying an external consultant to process it. For a business like 3Lines, and honestly for most organisations that still run on some volume of handwritten records, that’s not a small gap – it’s most of the paperwork that isn’t a polished final document.

Solution

Rather than a single fixed-scope build, we moved into an ongoing embedded engagement with 3Lines – one day a week – and prioritised OCR capability first, extending Regi so it can read handwritten notes and scanned documents in addition to the typed Statements of Advice it already reviewed.

• Deepening trust, not starting over – this is the same client, the same product, expanded because the original build delivered real value.
• MVP first, features layered on after – exactly the pattern we build toward with every client: get something working and trusted quickly, then keep adding capability as the relationship proves itself.
• Solves a genuinely underserved problem – most businesses still run on some volume of handwritten or scanned paperwork that AI tools simply can’t use; giving Regi that ability opens up a whole category of documents that were previously invisible to it.
• Embedded, ongoing delivery – a standing weekly engagement rather than a one-off project, so priorities can keep shifting as 3Lines’ needs evolve.

Result

The OCR feature is live, and it’s changed what handwritten and scanned documents cost 3Lines to process, in both time and money. Before, getting a handwritten note or scanned document into a reviewable, compliant format meant bringing in an external consultant – around 8 hours of work at $200 an hour, which is $1,600 per document, and often a wait before that consultant even had availability to start. Now the same document runs through Regi in around 10 minutes, for roughly $5 in AI credits – from 8 hours to 10 minutes, and from $1,600 to $5.

A final audit still happens on every document – that was already part of 3Lines’ process to guarantee compliance and accuracy, and it hasn’t gone anywhere. But the hours and cost that used to sit in front of that audit have effectively disappeared, and the wait for consultant availability is gone entirely.

Under the Hood