Agentic channels
People ask ChatGPT, Claude and Grok about life cover before they visit a website, and some send the assistant to get the quote. Subcontext lets those assistants reach you, with your answers and your prices, and records every conversation.
AI Ready Index · November 2026
Assistants are already shopping for insurance: ten million US users signed up to Meta's Muse within weeks of its launch, and Grok is live in the UK. So we gave one a real identity and a phone number and sent it to buy a policy. 3 firms blocked it, 85 had nothing to say to it and 12 gave it a price. None sold it a policy.
A firm an assistant can't reach never makes its shortlist. The AI Ready Index shows what one sees at each firm's door, how far it gets and where it stops. We run it every month.
Captions are in the film. Float Insure is Subcontext's showcase insurer, and prices shown are examples.
Inside ChatGPT and Claude
ChatGPT and Claude both let a company add itself as an app. Once yours is in, a person who asks about your products gets your answer in the chat, with your name on it: a live quote, whether they're eligible, a callback with your team.
Prices come from your own rates, and answers from what you've told us is true. One assistant speaks for you in both, and your dashboard shows which one each customer came through.
Here's a quote from YourBrand:
YourBrand can call you this week:
On your website
Some assistants browse the web for people. Told to "get me a quote from YourBrand", one opens your site and fills in the form itself, guessing what each box means. It often gets it wrong or gives up.
WebMCP lets your site tell it plainly what it can do: give a quote, explain the cover, book a callback. Websites and quote forms built on Subcontext support it, so the assistant asks properly and gets a proper answer.
Partner systems
Brokers and comparison sites are starting to run assistants of their own. Instead of a person emailing you for a quote, their software asks yours and gets an answer in seconds. The standard for this is A2A, short for agent to agent.
Each partner you work with gets its own way in. The same assistant that speaks for you everywhere else answers them, and every request is recorded and shown in your dashboard like a customer's.
Software asking software, with you in the middle.
You stay in control
A conversation inside ChatGPT can feel like it's happening somewhere you can't see. It isn't. Whichever way a question arrives, you get the same record as for your web chat:
Each answer is checked against your rules before it goes out. The assistant only says what you've told it is true and only offers what you sell. Anyone who seems vulnerable goes to a person, and you can switch it off the moment you want to. The compliance page has the detail.
Questions
ChatGPT and Claude, through an app in each. Assistants that browse the web for people can use your website, and WebMCP makes it readable to the ones that support it. Partners' own systems connect through A2A.
No. The setup is done from the Subcontext dashboard by whoever manages your products. The only technical-looking step is pasting a link into ChatGPT's or Claude's settings, and we show exactly where.
Yes. Quotes come from the same rates and rules your website uses. Change a price once and ChatGPT, Claude and your website all quote the new one.
It only says what you've told it is true about your products, every answer is checked against your rules before it goes out, and anything it can't answer goes to a person. You can read every conversation afterwards.
Yes, instantly, from the dashboard. It stops answering the moment you do.
Into your Subcontext inbox, like a web chat. The assistant carries the conversation; the quote and the customer's details are recorded on your side.
Once your app is listed in an assistant's public directory, yes. Listing is a separate step that goes through OpenAI or Anthropic. Until then you and anyone you share the link with can use it, which is how most firms start.
Our monthly test of UK protection websites: an AI assistant visits each one and tries to buy. It shows what the assistant saw at the door, how far it got and where it stopped. Ask us for your firm's results: they're private and free.
Getting started
There's nothing to build. If your products are already on Subcontext, it goes like this.
That gets your team using it straight away. Listing it publicly in an assistant's directory is a separate step with OpenAI or Anthropic, and we prepare everything they ask for.
Book a call and we'll set it up with you on the call, or start with your firm's AI Ready Index results.