Coverage Cat offers umbrella checkout through personal AI agents
Max Cho and Gabriel Botelho's five-state brokerage lets assistants prefill applications and follow offers toward payment, with shopper confirmation before submission.
By RuntimeWire Staff · Published
Primary source: Coverage Cat
Why it matters
Coverage Cat is making its brokerage available inside shoppers' existing AI assistants. Its documented umbrella flow reaches toward checkout, but state coverage and the quality of carrier offers will determine its usefulness.

Max Cho and Gabriel Botelho have built an umbrella insurance workflow for personal AI agents at Coverage Cat. The brokerage's public MCP configuration lets an assistant begin an application, work through corrections and follow a selected offer toward checkout while the shopper stays in a conversation. Its umbrella purchase documentation lists an April 17th, 2026 update, but the tools' initial release date is not established.
Cho, Coverage Cat's co-founder and CEO, came to insurance through a claim of his own. He has said he knocked out several teeth as a teenager and read his policy while trying to work out how to pay for their replacement, according to Y Combinator's profile. His subsequent work on marketing products at Google and financial prediction software at Two Sigma gave him a different way to approach the problem: use data to match coverage to a household's risks, rather than leave a shopper to choose from a list of premiums. Cho is also a licensed insurance broker.
Botelho, Coverage Cat's co-founder and CTO, builds the quoting infrastructure, carrier integrations and compliance systems behind that proposition. The two met while working on Yale's student newspaper, according to a 2023 founder interview. Botelho previously worked on data products at CompStak and founded Casetab, a legal-software venture that participated in the Y Combinator Fellowship. Their current bet puts software at the front of an insurance purchase while retaining a licensed brokerage behind it.
A quote begins before the questionnaire
The public MCP configuration directs a shopper's assistant to start an umbrella request with umbrella_consumer_prefill. The assistant can use information the shopper has already provided, then check Coverage Cat's staged review for missing or incorrect details. Coverage Cat instructs the assistant to tell the shopper that an application has not been submitted until the shopper confirms the final review. The documented consumer tools then cover offer selection, bind-stage follow-up and document attachment. Credit-check consent comes after an offer is chosen. The documented steps do not guarantee an immediate policy for every applicant.
Coverage Cat draws a consequential line between two kinds of access. Its default consumer MCP surface is described as requiring no operator credential; follow-up calls use an intake token returned when the draft is created. The separate developer portal and delegated operator tools serve integrations that have an operator credential. The consumer path documents umbrella checkout within the agent flow, while the homeowners path sends shoppers to Coverage Cat's portal for the final bind. For a developer deciding what an assistant can actually finish, that distinction matters more than the presence of an API.
The product also changes where Coverage Cat meets a potential customer. A shopper who begins with a personal assistant can reach Coverage Cat's brokerage without first filling out its website intake. In a 2023 account of the founders' plans, Cho described paid advertising and company outreach as routes to customers. Making the brokerage callable by agents offers another route, with less of the initial form-filling the founders set out to remove. The documentation does not establish whether agents have brought in customers.
The brokerage still has to win the comparison
Coverage Cat says it compares home, umbrella, auto and renters coverage, does not sell shopper leads and can show options even when it earns no commission on a policy. Its site says it has served "tens of thousands of users"; that is Coverage Cat's measure and does not establish how many bought policies or used the agent workflow. Coverage Cat lists shopper availability in California, Florida, New York, Texas and Washington. The five-state footprint limits where an agent can turn a conversation into an eligible application.
A conversational insurance purchase is already contested. In an August 18th announcement, Insurify said its ChatGPT integration could show eligible auto shoppers real-time quotes and, depending on the carrier, let them take the next step toward purchase in chat. Coverage Cat is pursuing a different entry point: an agent-facing interface for umbrella insurance that a shopper can use with a tool-capable assistant. The distinction may matter for households coordinating liability protection across cars, homes and other assets, but a convenient interface cannot by itself establish competitive prices or suitable coverage.
Coverage Cat has operated since 2022 and joined Y Combinator's Summer 2022 batch. A September 2022 securities filing recorded $1.64 million sold to 18 investors in an offering with a $10 million target. The target was an offering amount, not money raised or a valuation. Resolute Ventures lists Coverage Cat in its portfolio, though the filing does not identify which investors bought into that offering.
For Cho and Botelho, the agent workflow extends the idea behind Coverage Cat's original brokerage: collect the relevant household facts once, check the policy against the risk and keep a broker available when judgment is needed. The technical documentation shows how an assistant can advance an umbrella application. Whether that path yields better coverage at a better price will be decided by the offers shoppers receive and the policies they ultimately bind.