More than nine in 10 Australian brokerages have experimented with or adopted artificial intelligence. Yet only about one-third of those doing so have a formal AI policy in place, and more than one in five have no AI-specific safeguards at all.
That is among the central findings of Technology to Trust: AI, Technology and Broking, released by the National Insurance Brokers Association (NIBA) at the 2026 NIBA Convention on the Gold Coast. The report, produced in partnership with Steadfast Group and drawing on insights from CoreData, is based on a survey of brokers and brokerage leaders conducted in August 2026.
The timing matters. Both the Australian Securities and Investments Commission (ASIC) and the Australian Prudential Regulation Authority (APRA) have flagged AI governance as a regulatory priority this year, APRA has confirmed it will conduct targeted operational and technology risk reviews of general insurers during 2026-27, and new Privacy Act requirements taking effect in December 2026 will require organisations to disclose certain decisions made by computer programs, including AI.
While 91% of brokerages have engaged with AI, only around one in four have embedded it into core business operations. More than a third cannot identify who holds responsibility for AI within their organisation.
Among small and medium brokerages, human review of AI outputs is the most common safeguard, used by 49%. For large and enterprise brokerages, staff training on responsible AI use leads at 62%.
“Brokers are experimenting widely with AI. The next step is turning that experimentation into planned, well-governed adoption that improves outcomes for clients,” said NIBA CEO Richard Klipin.
For individual brokers and practice managers, the question is practical: if an AI tool produces an inaccurate coverage comparison or policy summary, who in the brokerage is accountable? For many firms, the NIBA data suggests that question has not been formally answered.
Some brokerages have begun putting structure around their AI use. Shane Brady, director of Stone Lane Broking & Risk Advisory, told Insurance Business his firm has documented which AI tools staff may use and how client information must be handled. “At Stone Lane we have clear expectations and boundaries around what the approved AI tools are that staff can use,” Brady said.
Cynthia Yap, managing director of Dynamic Insurance Services, said an approved-tools list alone is not sufficient. She recommended that brokerages check how AI is actually being used in practice and assess tools based on data storage, access, retention, and model training. “An approved list alone does not establish that visibility,” Yap said.
Read more: Brokers are using AI but many brokerages haven’t written the rules yet
NIBA’s findings align with what regulators have observed across the broader financial services sector. In its 2026 Key Issues Outlook, ASIC reported “variable maturity in how businesses manage AI governance risks.”
APRA went further. In an April 2026 letter to regulated entities, the prudential regulator said governance, risk management, and operational resilience are not keeping pace with AI adoption. Its 2026-27 Corporate Plan confirmed that entities using AI will need to demonstrate stronger governance, risk management, and board oversight, and that general insurers face targeted technology risk reviews in the year ahead.
For brokerages, the regulatory trajectory points in one direction. Firms without documented AI policies and clear accountability structures face growing exposure as supervisory expectations tighten.
Read more: APRA pushes insurers to narrow AI risk oversight gap
One of the report’s most actionable findings is the link between strategic planning and AI maturity.
Brokerages with a detailed and resourced technology strategy are almost four times more likely to have embedded AI than those without one, at 35% compared with 9%. Large and enterprise brokerages are more than twice as likely as small and medium brokerages to have embedded AI, at 39% versus 16%.
The data suggests that for small and mid-sized firms looking to move beyond experimentation, the first step is not a new tool but a documented plan.
Policy analysis and coverage comparison lead adoption, with 86% of brokerages using or exploring AI for those tasks. The common thread across the leading use cases is information processing: reviewing, comparing, and summarising large volumes of documentation while brokers retain oversight and final decision-making responsibility.
Time savings are measurable. Fifty-eight percent of AI users report saving at least one hour per week, and the most common use of that recovered time is client advice and relationships, cited by 26% of AI users.
Forty-nine percent of brokers cite concerns about accuracy and reliability as the primary obstacle to further AI adoption. Data privacy follows at 43%, and regulatory uncertainty and a lack of internal skills or knowledge are each cited by 23%.
Forty-one percent of respondents said they need more guidance on responsible AI use, calling for governance frameworks, privacy standards, regulatory clarity, and industry best practice.
“Trust is what clients come to brokers for, and how we use AI needs to strengthen it. They are seeing real efficiency gains and putting more of that time back into their clients. The questions now are about accuracy, privacy, and governance, and the profession needs to answer them together,” said NIBA president Nick Cook.
Klipin said NIBA would use the findings to support members with practical guidance on responsible AI use and to advocate for a regulatory environment that supports innovation while protecting clients.
