Industry Outlook: Insurance — Week of September 28, 2026

Market Outlook

• US mutuals rebound sharply on 2025 earnings. AM Best reports US property and casualty mutuals doubled net income to about $42.6 billion in 2025, with underwriting swinging from a $7.2 billion loss in 2024 to a $14.8 billion gain. That earnings recovery gives mutual carriers more room to fund claims automation, core modernization, and AI underwriting programs, and raises competitive pressure on stock carriers that lag on expense and loss ratio improvement. (Insurance Journal, Sep 25)
• Severe weather keeps property cat risk in focus. Heavy rains in South Florida and a nor’easter hitting the Northeast and Mid Atlantic, alongside the approach of Hurricane Nolo toward Hawaii, reinforce that flood, wind, and drainage failures remain front line loss drivers. Carriers with IoT flood sensors, parametric triggers for rainfall and wind, and automated catastrophe claims workflows will be better positioned as event frequency and secondary perils keep pressuring property portfolios. (Insurance Journal, Sep 25, Insurance Journal, Sep 25, Insurance Journal, Sep 25)
• State Farm expands claims workforce capacity. State Farm plans to increase its claims workforce by about 10 percent, or roughly 3,000 employees, through 2027, explicitly tying the move to investment in training, development, and recruiting. That hiring signal suggests large carriers expect sustained claims volume and complexity, and will mix human capacity with automation rather than betting on straight-through AI processing alone. (Insurance Journal, Sep 25)

Discussion: Expect continued capital for tech programs but under tighter performance scrutiny. Claims volume and climate risk trends argue for accelerating automation and parametric experiments while keeping human expertise central in complex cases.

Headwinds

• Regulators tighten transparency on underwriting decisions. The Texas Department of Insurance launched a public tool that shows consumers why carriers decline, non renew, or cancel home and auto policies, exposing granular decision reasons. That transparency will push carriers to clean up rule sets, document AI and ML models used in eligibility and pricing, and ensure adverse action explanations align with both legacy and algorithmic decision logic. (Insurance Journal, Sep 25)
• Cyber and AI misuse risks surface in public sector. OpenAI disclosed that its bots accessed public data from multiple US government agency sites during test exercises, and Australia used the UN stage to discuss an OpenAI related breach while pushing for controls on algorithms and smart devices. Even though the events center on a vendor, they highlight rising regulatory and public concern around AI agents crawling sensitive domains, which will shape how carriers are allowed to deploy generative AI in underwriting, claims, and customer portals. (BBC Business, Sep 26, BBC Business, Sep 24)
• Large fraud and abuse cases heighten compliance risk. Minnesota’s $18.5 million civil settlement over fraudulent child nutrition claims and fresh litigation over historic abuse at schools and youth organizations keep fraud, abuse, and long tail liability under the spotlight. Carriers need stronger fraud analytics across both indemnity and benefits lines, and more precise policy wording and archival of coverage positions for abuse and misconduct exposures. (Insurance Journal, Sep 25, Insurance Journal, Sep 25)

Discussion: Compliance, model governance, and AI security should be treated as first class product requirements. CTOs should assume regulators and plaintiffs will ask for decision traces, model documentation, and audit logs for both underwriting and claims systems.

Tailwinds

• Regulatory tools create data for underwriting analytics. The new Texas public database of declinations, non renewals, and cancellations for home and auto policies will generate a rich data source on market conduct and risk appetite across carriers. Analytics teams can mine this data to benchmark underwriting rules, detect competitive gaps, and support explainable AI models that mirror acceptable regulatory patterns. (Insurance Journal, Sep 25)
• Mutuals’ profit recovery funds modernization budgets. With US mutual insurers’ net income roughly doubling year over year and underwriting back in the black, boards will be more willing to back multi year modernization programs. That budget window favors investments in core replacement, IoT based risk models for property and auto, and parametric pilots for flood and wind that can reduce loss adjustment expense. (Insurance Journal, Sep 25)
• High profile cyber theft boosts demand for coverage. The reported $351.6 million theft from crypto exchange Bitget, along with a temporary suspension of withdrawals, will push digital asset firms and exchanges to reassess crime, cyber, and tech E&O coverage. Insurers that already built specialized cyber underwriting models for exchanges and custodians can expand market share, especially if they pair coverage with security posture assessments and incident response services. (Insurance Journal, Sep 25)

Discussion: Use the earnings window and market demand spike to fund data infrastructure and specialized products. CTOs should push for reusable components for regulatory data ingestion, cyber risk scoring, and parametric trigger processing rather than one off builds.

Tech Implications

• Claims automation must scale alongside human hiring. State Farm’s plan to add about 3,000 claims employees signals that even the largest carriers do not expect AI alone to handle rising event driven volume. Engineering leaders should design claims platforms where AI triage, document extraction, and fraud scoring support adjusters, with clear human in the loop controls for complex bodily injury, commercial, and catastrophe claims. (Insurance Journal, Sep 25)
• AI agents require strict access and audit controls. Reports that OpenAI bots accessed public data across US government agency sites during tests, combined with Australia’s public disclosure of an OpenAI related breach, highlight that autonomous or semi autonomous AI crawlers can cross sensitive boundaries. Insurance CTOs experimenting with AI agents for underwriting data gathering, claims investigation, or regulatory monitoring need strong allow lists, rate limits, and immutable audit trails to satisfy both security teams and regulators. (BBC Business, Sep 26, BBC Business, Sep 24)
• Climate events push IoT and parametric architectures. Recent flooding in South Florida, a major nor’easter in the Northeast, and Hurricane Nolo’s threat to Hawaii show recurring localized events that are well suited to sensor driven and parametric solutions. Architectures that combine third party weather feeds, on the ground IoT telemetry, and smart contract style trigger engines can shorten time to pay and reduce manual adjusting for defined event thresholds. (Insurance Journal, Sep 25, Insurance Journal, Sep 25, Insurance Journal, Sep 25)

Discussion: Prioritize architectures that make AI assistive, observable, and controllable. Event driven, API first designs that plug in weather, IoT, and external data while maintaining strong access control and auditability will age better than monolithic AI deployments.

CTO Action Items

First, sit down with compliance and product to review where AI and rules engines touch eligibility, declinations, and cancellations, and ensure you can produce regulator friendly explanations similar to what Texas is now publishing. Second, tighten AI security: inventory any autonomous agents, web crawlers, or external model integrations, and put them behind explicit allow lists, rate limits, and audit logging before regulators or customers ask. Third, use the current earnings and pricing window to lock in funding for data infrastructure that supports IoT ingestion, parametric triggers, and cyber risk scoring, since those capabilities align directly with emerging climate and cyber demand. Finally, align claims automation roadmaps with workforce planning so that new tools are built to augment the adjusters you will still be hiring rather than trying to replace them outright.