Summary
In this a16z conversation, partner Alex Rampell talks with Jean Deni, founder and CEO of Town and former CTO of Plaid, about why an AI-powered SMB tax startup was not worth pursuing as a venture-scale company, and how that led to building Town.
Deni explains that the tax automation product worked technically, and the business was healthy by pre-AI standards, with revenue just under $1 million in the first year and projections of $2.5 to $3 million in the second. But reaching millions of small businesses was a hard go-to-market problem, and incumbents could use AI to automate tax preparation for their existing customers faster than a newcomer could win share.
The discussion then turns to Town, an AI tool for knowledge-worker workflows such as meetings and meeting prep. A recurring theme is that the product experience cannot be just a text box. It has to show users what is possible and do the work for them. One example is a restaurant owner’s monthly supplier price check, which might cost about a dollar a month to run and find hundreds or thousands of dollars in savings in the one month it matters.
They also discuss model-agnostic routing, with the line that humans are slow routers, and a future with less human involvement in the loop.
📺 Source: a16z · Published October 19, 2026
🏷️ Format: Interview
