Optalitix adds Agentic AI to Insurance Underwriting Workbench

Optalitix, the London-based pricing and underwriting insurtech, has added agentic AI to its Optalitix Quote platform, letting underwriters at insurers, reinsurers and MGAs query their pricing systems in plain English rather than navigating dashboards, filters and reports. The functionality ships disabled by default and inherits the platform's existing role-based permissions, meaning only already-authorised users can reach it once a client switches it on.

The mechanic is a conversational layer that sits over the underlying platform. Instead of building a report to find out how many policies bound this week or which submissions are stuck awaiting sign-off, a user asks the question directly and the system answers. Optalitix cites examples such as summarising portfolio risk, identifying applications that need approval, and reallocating an absent underwriter's caseload across a team.

What can underwriters actually ask the system to do?

The initial use cases are operational rather than analytical: surfacing workflow status, summarising open cases, and redistributing work. Because the AI pulls answers through the same APIs and business rules that already run the platform, Optalitix positions the underlying system as the single source of truth, rather than a separately trained model that could drift from the live data. In an underwriting context, where a pricing figure has to be reproducible and defensible, that architectural choice matters more than the natural-language interface itself.

The rollout is phased. Optalitix says the first stage enables connectivity for clients' own external AI systems, followed by AI-generated quote summaries, workload management, broker-communication assistance and client-side co-pilots that interact with the platform through standard interfaces. A later stage targets portfolio-level decision intelligence: portfolio scoring, pricing benchmarking, underwriter co-pilots, automated management reporting, risk recommendations and explainability tooling.

Why does governance decide whether this gets adopted?

For regulated insurers, the constraint on AI has rarely been capability. It is auditability. Optalitix says the implementation follows its ISO 27001-certified information security framework, and that every action remains logged against the platform's existing audit trails and access controls. That is the crux of the pitch: adoption depends less on whether the AI is clever and more on whether a compliance officer can trace how it reached an answer and confirm the user was permitted to see the data behind it.

The timing tracks a broader tightening of model-risk expectations. Canada's OSFI Guideline E-23 brings AI and machine-learning models formally into model-risk-management scope from May 2027, and European supervisors are moving in parallel. Vendors that can layer AI onto an already-governed, version-controlled system have a cleaner regulatory story than those bolting analytics onto legacy spreadsheets, which is the environment much of the mid-market still runs on.

How does this fit the current insurance market?

The launch lands in a softening market, where rates are falling and carriers are under pressure to protect margins without the tailwind of hardening prices. In that environment, the value case for tooling shifts from writing more premium to extracting more from the data insurers already hold: tighter pricing consistency, faster case handling and lower operational cost per policy. Founding director Dani Katz framed the market backdrop as one where organisations increasingly treat AI as a competitive requirement rather than an option, and where value depends on the underlying pricing system being fully digitised first.

Optalitix converts pricing models built in Excel, Python and R into cloud-hosted APIs, and counts recognition from the Lloyd's Lab, MassChallenge and the Oxbow Partners InsurTech Impact 25 among its credentials. The company recently signed Dutch MGA Intermont, part of the Acrisure group, to modernise its pricing operations, indicating continued expansion beyond its UK base.

The agentic AI move places Optalitix alongside a widening field of vendors embedding conversational and generative AI into underwriting workflows, where the differentiator is increasingly not the model but the governance wrapped around it.

Why This Matters to FinanceX Readers

For investors and finance professionals tracking insurance technology, the signal here is that AI in underwriting is consolidating around governance rather than raw capability. In a soft market, carriers that can compress case-handling time and enforce pricing consistency defend combined ratios without needing rate increases, which turns operational tooling into a margin lever.

The vendors likely to win enterprise deployments are those, like Optalitix, whose AI inherits existing audit trails and access controls, because that is what lets a regulated insurer say yes ahead of tightening model-risk regimes such as OSFI's E-23. The practical question for anyone holding insurer or insurtech exposure is which platforms can turn natural-language access into measurable underwriting leverage rather than a demo-stage feature.