Onboarding, identity checks, document chasing, file notes, review cycles and commission reconciliation, built so the record is a byproduct of the work.
In a broking, advice or accounting practice, the file is the product. The advice is the value, but what you are actually paid for — and what you get reviewed on — is a complete, dated, attributable record of what was recommended, what was disclosed, what the client agreed to and when. Custom software for financial services in Australia earns its place when the record has stopped being a byproduct of doing the work and started being a second job somebody does on Thursday afternoons.
This page is for the intermediary end of the market. General insurance and mortgage brokers, financial advisers and authorised representatives, accounting and bookkeeping practices, claims and premium funding businesses, small licensees. Firms of five to eighty people, running a core platform they did not design and cannot alter — Xplan, AdviserLogic, Insight, WinBEAT, an aggregator CRM, Xero Practice Manager, MYOB — plus a spreadsheet, a shared mailbox and a set of habits holding the gaps together.
We are not a licensee, not a compliance consultancy and not a licensed adviser. We do not interpret the Privacy Act, tell you what your AFSL or credit licence conditions require, decide your record retention periods, or sign off that a system makes you compliant. Those are questions for your responsible manager, your licensee, your industry association and your lawyers. We build the system that captures the evidence they tell you to hold, and makes it retrievable in seconds rather than an afternoon.
What tends to make the difference in this sector is not a new platform. It is the connective work around the one you already run. That is system integration and, in most practices, a good deal of document AI — because the inputs arrive as PDFs, scans and phone photos, and always will.
One onboarding path that collects what you need in the order you need it, records which identity documents were sighted and verified, by whom and on what date, and stops a file progressing while something mandatory is missing. Where AML/CTF or verification of identity obligations apply to you, your adviser defines the requirement and the system holds the proof it was met.
The unglamorous one that pays for itself. Outstanding-items lists per client, a secure upload link instead of an email attachment, automated reminders that stop the moment the document lands, and a clear view of which files are waiting on the client versus waiting on you. Personal information stops travelling through individual inboxes.
Structured file notes captured at the time, tied to the client, the engagement and the person who wrote them. Templates for the record types you actually produce — advice records, renewal discussions, claims file notes, scope changes — with an immutable history of who entered what, who changed it, and when. That history is what separates a record from evidence.
Renewals, annual reviews, fee consent and disclosure cycles, BAS and lodgement deadlines run as a tracked pipeline rather than a spreadsheet somebody sorts by date each month. Work is allocated, escalates when it stalls, and shows plainly what is due, what is overdue and what has been actioned but not evidenced.
Insurer, lender and platform statements arrive as PDFs and CSVs in a dozen formats. They get read, matched to the policy, loan or client, and exceptions surfaced — underpaid, missing, clawed back, paid on a client who left two years ago. Most practices know they lose some revenue here and cannot quantify it, because checking every line by hand is not realistic.
Clients see their own documents, policy schedules, certificates of currency, statements and forms behind a login, and return signed items to the same place. Access is enforced per record, not per screen. We cover the mechanics on our client and customer portals page.
We work through what you have to be able to evidence, to whom, and how often. Licensee requirements, client agreements and insurer or aggregator conditions usually generate more obligations than the regulator does, and they are the ones nobody has mapped.
We document how each record is created today, who touches it, and where it comes to rest. This is normally where a practice finds the same client details entered three times and the one item a reviewer asks for never captured at all.
We agree what the system holds, who can see it, what escalates and when. Access design is deliberate here — record-level permissions, not hidden menus — and so is retention, which you specify and we implement.
We build it and connect it to your practice management, accounting and email systems so people keep working where they already work. Client data we host sits in Australian regions, on Azure in Melbourne and AWS Bedrock in Melbourne, which removes an awkward question from most licensee and insurer due diligence.
Once it is running we look at where files stall, which reminders get ignored and which fields go unfilled. Those are design problems, and they are cheap to fix once visible.
No to both, and we will not drift into it. Drawn AI builds software. We do not hold an AFSL or a credit licence, we do not provide legal, tax or compliance advice, and we will not tell you what your obligations are or certify that a system meets them. You or your advisers define the requirement. We build the system that captures and produces the evidence.
In Australia. Client data we host sits in Australian regions — Azure Melbourne and AWS Bedrock Melbourne. Access is enforced at the record level, every read and change is logged, and your team controls who has access and can revoke it immediately. The data is yours. If the subscription ends, you get it out in a usable format.
Usually, but the shape changes. Where a licensee requires advice to be produced and stored in a nominated platform, that stays as the system of record, and what we build sits around it — onboarding, chasing, reconciliation, portals. Confirm the approach with your licensee before you commit to anything. Some are specific about what may sit alongside their platform.
That is one of the better uses of document AI in this sector. Statements arrive in inconsistent formats, and a system can read them, match lines to policies or loans, and present only the exceptions for a human to judge. Accuracy is never perfect, so we build it to flag uncertainty rather than quietly guess, and a person approves anything that does not match cleanly.
We work on a subscription rather than a fixed project price — monthly billing, a three-month minimum term, thirty days' notice to cancel. The figure depends on how many people use it, how many systems it connects to, and how much document handling is involved. Email outreach@drawnai.app with a description of your practice and we will explain what drives the number.
Do not start one in the middle of renewal season or the run-up to a lodgement deadline. Expect a few hours a week from someone senior during discovery and design, because they hold the knowledge of how files really move. Rollout is staged — one process, one team — so nobody is learning a new system and clearing a backlog at the same time.
A first conversation costs nothing and usually ends with a clearer idea of what is worth building — sometimes that answer is “not yet”, and we will say so.
No lock-in templates. Built around you.