Discussion

i23 raises £4.9m to put AI agents to work in insurance broking

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#9857

Former UK Treasury minister Bim Afolami has co-founded i23, an AI start-up that has raised £4.9m to automate routine work in insurance broking. The company also plans to buy an existing broker, a move that could give its software a live business to work in rather than a theoretical workflow to tidy up.

Watch Desk analysis

What happened

i23 is building AI agents to handle brokers’ routine administration. Co-founder Anand Sambasivan is chief executive; Afolami, who served as economic secretary to the Treasury under Rishi Sunak, will be president. Sky News reported the seed round on 28 October, with Seedcamp, Outward VC and Insiders VC among the investors named.

The company expects to acquire an insurance broker in the coming months. Sambasivan says the aim is to help brokers serve clients better and grow revenue without growing administrative work at the same pace. That is the company’s pitch, not yet a demonstrated result.

Why it matters

Buying a broker would give i23 access to real clients, policies and paperwork, and let it test its agents inside the business it hopes to improve. That is a more direct trial than selling software to brokers one by one, but it also means the company must change how people work, not merely install a new tool.

Insurance broking is a large, fragmented market, according to Sky’s report. If i23 can show that automation improves a broker’s results without making service worse, the approach could attract interest well beyond its first acquisition.

Our read

The funding is useful; the planned broker purchase is the part to watch. AI agents can make a convincing pitch for repetitive administration, but broking involves policies, clients and judgement, not just forms waiting to be filled. The proof will be whether the business works better after the software arrives.

What to watch

  • Whether i23 completes its planned broker acquisition, and when.
  • Which administrative tasks its agents actually handle in the acquired business.
  • Whether i23 publishes evidence of improved service, revenue or operating costs.
  • How much human review remains necessary when an agent makes a mistake.

Discussion spark: Should AI companies buy the businesses they want to automate so they can prove the model in practice, or does that make it harder to judge their software independently?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.