Subcontext, an artificial-intelligence-native insurance business, has launched a platform that enables insurers and brokers to distribute protection products directly through large language model environments such as ChatGPT and Claude, and via messaging channels including WhatsApp.
What the technology does
The company says its system lets customers discuss protection needs with conversational AI assistants, receive indicative pricing, check eligibility, review exclusions and request follow-up contact without leaving the chat environment. Subcontext has also integrated with Iress’ The Exchange platform, enabling whole-of-market distributors to access insurer panels through the same technology.
“Consumers who do not want to buy protection over the phone are underserved by today’s protection market because traditional distribution cannot profitably deliver a quick, easy purchase,” said Alain Desmier, co‑founder of Subcontext.
Mike Laming, co‑founder and chief technology officer, described the product as "the agentic insurance stack" that operates on top of large language model infrastructure and includes governance controls, adding that journeys can appear within an insurer’s own brand or through messaging channels.
Why insurers are interested
The proposition seeks to meet customers who favour digital interactions rather than telephone sales. Subcontext frames conversational AI as a way to open new distribution channels on the platforms consumers already use, and to automate parts of a sales and eligibility process that traditional distribution finds hard to deliver profitably at scale.
The launch follows the Financial Conduct Authority’s review of AI in retail financial services published in July 2026, in which the regulator identified AI as a potential tool to broaden access to protection products while flagging security and oversight issues. The FCA urged firms and supervisors to secure the regulatory perimeter, strengthen co‑ordination and oversight, and monitor the shift to more autonomous models.
Practical implications for customers, prices and jobs
Subcontext’s platform could make it easier for consumers who prefer text or in‑app interactions to explore protection options without a phone call. That convenience may increase take‑up among digitally engaged groups, but the effect on pricing is less certain.
- Insurers may be able to reduce distribution costs per sale if AI handles initial engagement, which could lower the marginal cost of selling a policy.
- Automated conversations could speed eligibility checks and comparison across panels when integrated with platforms such as The Exchange, potentially improving price transparency.
- There are employment implications for telephone sales and contact centres if routine conversations migrate to automated channels; the FCA has advised monitoring transitions to autonomous models.
None of these outcomes is guaranteed. Firms still face responsibilities under conduct rules to ensure customers understand products and exclusions. The FCA’s recommendations suggest that regulators expect firms to manage risks such as data security, model governance and the potential for algorithmic bias.
| Feature | Promised benefit |
|---|---|
| Conversational AI in LLMs | In‑conversation pricing, eligibility and product explanation |
| Integration with Iress’ The Exchange | Access to whole‑of‑market insurer panels |
| Messaging channel deployment | Sales via platforms consumers already use (eg WhatsApp) |
Regulatory and industry scrutiny
The FCA’s July 2026 review acknowledged the potential for AI to broaden access to protection but set out several security and governance recommendations. Firms deploying conversational AI will need to show how they maintain oversight, protect data and ensure fair treatment of customers — especially where decisions become more automated. The regulator will be watching how firms adapt the existing regulatory perimeter to AI‑driven sales.
For insurers and distributors, the technical integration is only one hurdle. They must also be able to demonstrate that automated interactions meet suitability, disclosure and record‑keeping standards and that any fallbacks to human advisers are timely and effective.
Subcontext’s offering is a timely example of how generative AI is being built into financial distribution. The next phase will test whether these tools can combine customer convenience with the safeguards regulators demand, and what that balance means for the cost and accessibility of retail protection and the employees who handle sales and advice today.
