AI agents are powerful, but most ordinary people don’t yet use them. Town wants to change that. Led by CEO Jean-Denis Greze, playfully referred to as Town’s mayor, the company announced a $108 million Series B on October 19, co-led by Forerunner Ventures and Menlo Ventures, at a $1 billion valuation.
Greze cofounded the San Francisco startup with Tony Vincent in late 2024 after serving as CTO of Plaid, where he built up the engineering team from about 20 people to 350.
Town’s betting it can make AI more accessible and practical for small-business owners and mid-market firms still looking for day-to-day AI benefits.
Fast Company spoke with Greze about why most people still treat AI like a fancy search engine, where AI assistants can be most helpful, and why he’s working harder than he ever has before. The conversation has been edited for length and clarity.
ChatGPT and Claude can already connect to other apps. Why do people need stand-alone AI assistants like yours?
Could Claude or ChatGPT do most of what Town does? I think the answer is, most of it, yeah, they can. But most people who use ChatGPT or Claude aren’t using those tools to do what people in tech describe as agentic work, which just means using them to actually do things for you as opposed to researching things for you. They use it more as a Google Search replacement.
I think the biggest reason is that most people don’t really know what to put in the chat box. Putting someone in front of an empty box and asking them to write what to do, especially when the capabilities are changing all the time—that’s not onboarding.
No one’s figured out how you take a mainstream person and get them to use these products to their full potential. That’s the market opportunity. And we may be wrong, by the way. Maybe distribution just wins, and Google, Microsoft, and OpenAI end up winning. As a startup, that’s the risk you’re taking on.
What’s a use case that surprised you?
There’s a small travel firm at a vacation destination where people rent cars and gear, like a bike, for the day. The business is totally run over email. If something’s damaged, the employee takes pictures and sends an email to the central processing person.
Now there’s a little workflow that at the end of the day takes all the emails with the pictures, looks at the CRM for who the people are, and drafts the emails. They were drafting 10 to 12 of these emails every day, and it wasted 25 minutes; now it takes them 5 minutes. No one’s building SaaS for that small business.
What should we call this category of tool? An AI assistant? A chief of staff?
We call our product a Townie. If you use the word assistant or the phrase chief of staff , you have a human mental model for what an assistant does. I think what this technology does today and what it’ll do in five years will probably defy these expectations. And no one knows what a chief of staff does outside of CEOs of Fortune 500 companies.
I actually think the right analogy for what we’re building is more like a browser or a device. When the App Store came out, suddenly a percentage of what you used to do in your web browser, you started to just do in your phone. Adding AI assistance is the same.
The percentage of the time you interact with information in a browser or your phone will just go down now forever, and the percentage that you do by talking to an AI will go up. The future, I think, is voice for sure.
What’s next for Town?
I’m very excited about live voice. You could imagine in the morning you’re commuting to work and you have in your ear a world-class executive assistant that’s telling you, “Hey, you got seven emails overnight. There’s two of them that are urgent. On this one, if you tell me this, I can help draft the reply. The other one is going to require you to move a meeting later today. Do you want me to come up with some options?” It’s just talking to you, and you can interrupt it.
I’m also excited about what we call agent to agent. You could say, “Who at the company is really close with someone doing R&D at this other company?” Your agent will talk to other agents at your own company, and it’ll learn that Bob plays tennis with someone there. But it won’t tell you that. It’ll just be like, “Hey, Bob appears to be closely connected to that person. Do you want me to ask them if they can do the intro?” Leveraging the network of everyone within a company in a way that’s privacy-preserving is huge.
Should people connect their financial accounts to an AI agent?
If you’re a consumer, it is very likely that you connect everything about you, your medicals, your financials, to the agents. It’s just too beneficial. It’ll tell you that your credit card rate’s too high, that you should apply for a different one. It’s free money.
But the privacy concern is real. I don’t think you want to use an assistant where you’re not convinced it’s making decisions that are in your best interest. These products only work in a world where you think they truly have your best interest.
Can you trust those services?
If you’re paying for it, you can. If you’re not paying for a service, you’re not the customer. You’re being monetized in some other way.
We made the choice very early: We like consumer use cases, and they’re subsidized by business use cases. We charge businesses, and it’s not a cheap product; it’s not $5 a month. Our main plan is $50 a month.
What’s the hardest part of running Town?
The world’s too fast. I’ve never seen anything like it. It’s nerve-racking. Software had high gross margins because there was a really high fixed cost to copying it, and there was a real advantage to being early in market. By the time someone could copy you, you’d have built the next feature.
With AI, competition moves so fast and good ideas are copied so quickly. Things that would take nine months to copy are copied in a month or two.
So the pressure you feel to execute and to work a lot—I’m 47, but I’ve not worked this hard [before]. Psychologically, it’s tough, because you’re either ecstatic because you shipped something before OpenAI did, or you’re in fear that there’s a tweet from Sam Altman that puts you on your back foot. So that’s super stressful.
How do you use Town outside of work?
I’ve told Town to research food blogs and newspapers that write about food in San Francisco. I have a date night with my wife every Sunday night. So every Monday night, it looks at the next month and gives me four tables that I can make a reservation at that are classics, and then finds four that are new that have been mentioned in food blogs. I like food. I just don’t have the time [to do the research myself].
