Insurance
The data you handle is exactly the data an underwriter would kill to see.
Broking runs on detail. To place a risk properly you gather claim histories, financial position, medical disclosures, sometimes deeply personal circumstances the client would not casually mention twice. Staff use AI to move that work along — summarising submissions, redrafting placement slips, comparing wordings. The material passing through the tool is precisely the material an insurer, a competitor, or a plaintiff’s lawyer would benefit from reading.
Request an assessmentYour obligation
ASIC, ACL and privacy overlap here in an awkward way
Brokers hold personal information and are subject to the Privacy Act and APPs — including APP 8 on cross-border disclosure. Beyond that sit AFS licence obligations, general insurance code obligations around handling information about vulnerable clients, and long-standing duties of good faith and confidentiality that predate anyone thinking about AI. Retail insurance product recommendations must be documented against the client’s relevant personal circumstances, which is the same detail that makes an AI request useful and a disclosure serious.
Where it’s happening
The four places we see it most
None of these are careless people. They are competent staff doing the work faster, using a tool that is genuinely good at it.
Placement submissions
Full risk detail assembled for the market — pasted in to tighten the summary, transmitting the entire commercial position of the insured in one request.
Claims correspondence
Insurer position, adjuster reports and client responses handed to AI for drafting help, carrying loss detail and any admissions with them.
Wordings comparison
Policy documents uploaded to explain differences to a client, along with the schedule identifying the insured and the specific cover.
Medical disclosures in life and income protection
Sensitive information under the Privacy Act — supplied so the AI can produce clean file notes, which is exactly the material that requires the highest standard of protection.
The alternative
Same capability, nothing leaves
A private system does not ask your team to give anything up. It does the same work on the same documents — it simply does it on a machine you own, sitting in your own office.
See how it works →- Summarise complex placement submissions without transmitting the risk detail
- Draft claims correspondence against the actual file, not a sanitised version
- Compare policy wordings against internal notes and prior placements
- Handle medical and financial disclosures without a cross-border transfer
- Give the whole broking team AI at a fixed cost, without training a variable habit
Insurance Brokers: common questions
- Is a broker actually liable for what a downstream AI service does with client data?
- Under APP 8 and section 16C, disclosing personal information to an overseas recipient generally keeps the discloser accountable for what happens to it. A broker cannot outsource privacy accountability to a chatbot vendor by pasting a submission into it, and the FSCP would not treat "an employee did it without telling us" as a defence to a complaint about disclosure.
- Does this help with insurer portals and platforms we already use?
- Platforms you have properly onboarded — Sunrise, iBAIS, WinBEAT, insurer portals — are known vendors with defined agreements. That is a managed risk. The unmanaged risk sits in ad-hoc AI usage that no one signed off on. On-premise AI closes the second gap without touching the first.
No obligation
Is this worth it for a firm like yours?
Tell us roughly how many people, what they handle, and what you suspect is already happening. We will give you a straight answer — including if the answer is that you do not need us.