Key Takeaways

• Insurance Australia Group (ASX:IAG) is scaling AI across its claims value chain, with more than 90 generative AI use cases live and over 2,000 staff using 'composed' AI in claims, fraud and service teams.

• Current tools such as CASI (a CGU claims assistant) and the Codafication Crunchwork property-claims platform speed up processing but keep human decision-makers in control rather than fully automating payouts.

• COO Neil Morgan has flagged CTP claims assessment and agentic AI as future 'channels to market', signalling ambition to automate more of the claims journey over time.

• IAG's H1 FY26 result showed GWP up 6.0% to A$8,929 million but reported net profit down 35.1% to A$505 million after severe weather; FY26 guidance was maintained.

• Investors should weigh AI-driven efficiency gains against regulatory scrutiny from APRA and ASIC, algorithmic fairness risks, weather volatility and a share price near 52-week lows.

Insurance Australia Group (ASX:IAG), the parent of household brands NRMA Insurance and CGU, is quietly rewiring one of the most labour-intensive parts of its business: the claim. Over the past two years the country's largest general insurer has moved from experimenting with artificial intelligence to embedding it across quote ingestion, claims lodgement, fraud detection and customer service, and management has begun openly discussing a future in which more of the assessment and settlement process is handled by machines rather than people.

That raises a provocative question. If an AI can already read a policy document in seconds and triage a property claim across a supplier network, how far is IAG from letting software decide and pay a claim on its own? Based on what the company has disclosed, it is exploring the territory carefully rather than flicking a switch. The tools live today assist human handlers rather than replace them, but the direction of travel is unmistakable.

What Has Happened

IAG has been steadily deepening its use of AI across operations, describing a three-part approach: "deploy" (AI built into vendor platforms such as Microsoft, Google, Guidewire, Earnix and ServiceNow, which around 60% of its workforce now touches), "shape" (certified internal "activators" who tailor AI to business needs), and "compose" (building more complex, at-scale solutions).

The numbers give a sense of the pace. IAG has moved from effectively zero generative AI use cases two years ago to more than 90 in production, with its pool of certified activators quadrupling since late 2024 to over 600. Crucially, the company says more than 2,000 of its people now use "composed" AI in everyday work — explicitly including claims, fraud and service teams.

Two tools illustrate what this looks like on the ground. The first is CASI — Claims Assistant Supporting Intermediated — an app that sits on a CGU claims handler's computer and answers natural-language questions about complex Product Disclosure Statements. According to Chief Operating Officer Neil Morgan, it has compressed the task of finding relevant policy terms from "20 to 30 minutes to seconds", while providing source references so a human can verify the answer. CASI was rolled out to roughly 500 front-line claims staff in CGU's intermediated division and is now expanding.

The second is Crunchwork, a platform from Australian software firm Codafication that IAG adopted for property claims. Announced in late January 2026, it connects insurers, customers, repairers and assessors on a single system, replacing fragmented phone-and-email workflows across the NRMA Insurance, CGU and WFI brands.

Why It Matters

Claims handling is where insurers win or lose customers, and where a large share of costs sit. Faster, more consistent processing can lift retention, cut expenses and reduce the leakage that comes from human error or fraud. For a group underwriting more than A$17 billion of premium a year across Australia and New Zealand, even small efficiency gains translate into meaningful dollars.

The strategic significance is bigger than cost, though. Morgan has publicly identified CTP (compulsory third party) claims assessment as one area ripe for AI, and has spoken about "agentic" AI — software that carries out multi-step tasks with limited human intervention — as potentially "another channel to market". That points toward a model in which straightforward claims could one day be assessed, and perhaps settled, with far less manual touch — a structural shift for a market long characterised by slow, paperwork-heavy claims.

AI is also a double-edged theme for investors: the same technology that promises efficiency has fuelled anxiety that insurers' competitive moats — pricing skill, claims expertise, distribution — could be commoditised, and IAG shares have not been immune to that debate.

How IAG Is Using AI Across the Claims Value Chain

It is worth being precise about what is automated today. Neither CASI nor Crunchwork settles claims by itself: CASI is an information-retrieval assistant that keeps a human handler in the decision seat, while Crunchwork automates workflow and data-sharing between parties but leaves repair and settlement decisions to people.

What IAG is really building is a layered capability. At the base, deployed AI within core systems — including Guidewire for claims management — standardises processes. On top, composed and agentic tools progressively take on repetitive cognitive work: reading documents, ingesting quotes, spotting anomalies, updating stakeholders. Full automated payouts would sit at the far end of that spectrum, and IAG has framed any move that way as requiring "appropriate human oversight and governance controls". The plumbing that could support automated settlement is being laid, but the company is signalling caution about switching off human judgement, particularly for contentious or high-value claims.

Key Financial and Operational Details

IAG's AI push is unfolding against a mixed financial backdrop. In its half-year result to 31 December 2025 (H1 FY26), the group reported gross written premium up 6.0% to A$8,929 million, helped by around A$500 million from RACQ Insurance (RACQI), acquired in late 2024. Underlying insurance profit rose 7.6% to A$804 million, for an underlying margin of 15.1%.

The reported numbers were weaker. Reported insurance profit fell 24.3% to A$724 million and the reported margin dropped 590 basis points to 13.5%, dragged down by severe weather. Net profit after tax fell 35.1% to A$505 million. Despite that, the board held the interim dividend at 12 cents per share, announced a A$200 million buyback, and maintained FY26 guidance of A$1,500–1,750 million reported insurance profit, a 14–16% margin, high single-digit GWP growth and a roughly 15% return on equity.

The share price has reflected the pressure. IAG touched a 52-week low around A$6.71 in February 2026, down close to 16% year-to-date, weighed by claims inflation, lower investment income and the ACCC's December 2025 decision blocking its proposed RAC Insurance acquisition.

Potential Catalysts

Several developments could move the story from here. Tangible evidence that AI is lowering the claims expense ratio or improving retention would validate the investment, as would expansion of tools like CASI beyond CGU into the larger NRMA Insurance personal-lines book. A credible pilot of more automated assessment in a contained area such as CTP or simple motor claims would be a proof point for the agentic ambition.

On the fundamentals, successful integration of RACQI, continued high single-digit top-line growth and a benign second-half weather experience would help the reported margin recover toward guidance, while continued buybacks and dividends could support sentiment as the AI narrative plays out.

Material Risks and Uncertainties

The risks are real and, on the AI front, partly novel. Regulators are watching closely: APRA issued guidance in April 2026 warning that governance, risk management and assurance are "not keeping pace" with AI adoption in financial services, flagging board capability gaps and over-reliance on single AI or cloud providers, while ASIC and APRA have jointly called for stronger AI oversight. Any move toward automated claims decisions would invite scrutiny over algorithmic fairness, bias, transparency and the risk of wrongly declined or underpaid claims — conduct hazards that could outweigh efficiency gains if handled poorly.

Beyond AI, IAG carries the classic general-insurance risks: natural-catastrophe volatility (already evident in the H1 result), reinsurance costs, claims and repair-cost inflation, investment-market swings and pricing-regulation scrutiny. Execution risk on RACQI integration and the setback from the blocked RAC deal add uncertainty, and AI could itself compress industry moats if competitors match IAG's capabilities.

What Investors Should Watch Next

For Insurance Australia Group (ASX:IAG), the near-term signposts are clear. First, watch for hard metrics — claims cost ratios, processing times, retention — that show AI is delivering, not just being deployed. Second, look for any formal pilot of automated or semi-automated claims assessment, and the governance framework attached to it. Third, monitor the reported insurance margin against the 14–16% FY26 guidance, since weather and claims inflation will determine whether underlying gains flow through.

Investors should also track regulatory signals from APRA and ASIC, RACQI integration progress and capital management. The full-year FY26 result will be the next major test of both the financial trajectory and the AI story.