Town Hits a $1 Billion Valuation by Letting AI Assistants Work as a Team

For the most part, an AI assistant is a one-on-one affair these days. It drafts your email, tidies up your calendar and gives you a rundown of your documents, then stops at the edge of your inbox. Town, a San Francisco startup, has put its money on the idea that the next step up in productivity comes when those assistants can work with one another.

To that end, Town announced on Monday a $108 million Series B round that values the company at $1 billion, only four months after its public launch. The cash came with new features that let an employee's assistant put questions to a colleague's assistant, pass along work and deal with people outside the company. It is a plain sign that Silicon Valley investors now see the AI assistant as shared workplace infrastructure rather than a personal tool.

What Town Announced

Forerunner Ventures and Menlo Ventures co-led the financing, joined by Better Tomorrow Ventures, Scribble Ventures, Serena Ventures and Sparked Ventures, as well as existing backers Andreessen Horowitz and First Round Capital. A group of angel investors took part too, among them Scott Belsky, Romain Huet, Brendan Foody and Josh Silverman.

Town launched publicly in June with a $55 million Series A, bringing its total funding to more than $160 million. The company says annual recurring revenue reached seven figures within weeks of launch, and that since July its paying teams have grown ninefold and its paying customers fourfold. It says thousands of businesses use its assistants every day, though those numbers come from Town itself and have not been independently verified.

How Townies Work

Every employee gets an assistant from Town, called a Townie. It has access to email, documents, calendars and the core systems a person relies on at work, and over time it picks up on that individual's projects, relationships and way of writing. The idea is an assistant that can shoulder some of the workload, not merely field questions.

"More people are trusting AI assistants with real tasks, and work is where they need the most help," co-founder and CEO Jean-Denis Greze said in the company's announcement. Greze was chief technology officer at the fintech company Plaid before he and co-founder Tony Vincent started Town.

From One Assistant to a Team of Them

The new release rests on a simple observation. "AI made each of us faster on our own, which makes working together the new bottleneck," Greze and Vincent wrote in a blog post on the round. Work can still come to a standstill while people wait on colleagues for data, input or approvals.

To address this, Townies can now talk to one another. An assistant can put a question to a coworker's Townie or hand off a task, but it remains answerable only to its own person, shares no more than that person would and asks permission when required. In the company's examples, a Townie might draft a message in an executive's voice for the executive to send, arrange an introduction through a colleague who knows a sales prospect, or run a presentation past teammates' Townies to get ahead of their questions.

Teams can also share workflows. A sales manager who adds a new field to the team's call prep does it once, and every rep's Townie picks it up. And Townies can now work over email and text with people outside the company, whether customers, partners or job candidates.

The Data Question

With an assistant embedded in every user's email and documents, one is bound to ask who controls what it learns. Town's position is that the employer owns the knowledge. "We believe that this context is a critical asset and should belong to the company that built it," the founders wrote, adding that it should stay with the company and be portable.

The company says that, by default, nothing a customer connects is used to train AI models. It holds SOC 2 Type 2 certification and publishes its data processing agreement. Rather than tie a customer to one provider, it calls on models from several large AI labs, choosing one for the job at hand. That multi-model strategy is becoming common among startups that want whichever option is best or cheapest at the moment.

What It Could Look Like at Work

Take a sales team on a Monday morning. A rep's Townie is getting ready for a client call and needs the latest pricing exception, so it asks the finance lead's Townie instead of waiting on an email reply. The manager has decided that all call prep should now include a competitor comparison, and the change reaches every rep's assistant in one go. When the call is done, the Townie writes up a follow-up and, after running it by the rep, sends it to the customer.

Individually, none of it is remarkable. The argument is that it removes many small delays over the course of a day. How well that works in reality depends on how often the software gets things right and how comfortable people are with it doing the talking.

Will Every Coworker Soon Have an AI Assistant That Talks to Yours?

The potential is considerable. With assistants able to work in concert, much of the day-to-day office routine could shrink: chasing a colleague for a number, wrangling five calendars to set a meeting, preparing for a presentation or bringing a team up to speed on a new process. For a small Silicon Valley startup, that could give a 20-person team the reach of a far bigger company. Town's investors are clearly betting the category will be big; as Menlo Ventures partner Shawn Carolan put it in the announcement, "In 2026, the technology has arrived, and the opportunity is bigger than ever."

Then there are the very real concerns. An assistant that can read email and talk to its peers opens up fresh avenues for error and abuse. If a Townie shares the wrong detail, or is duped by an artfully worded message from outside the company, it could spread information faster than any person could. There are also questions of trust and monitoring: employees may not like the idea that their assistant's knowledge of their relationships and habits belongs to the employer. And there is the matter of jobs, since the tasks assistants coordinate among themselves are often the ones that once filled junior roles. Regulators in California and Europe are already looking more closely at automated decisions in the workplace, which could add compliance costs for any company deploying these tools widely.

There is also a debate over who should win this market. Microsoft, Google and other large platforms already sit inside most companies' email and documents and are building agents of their own. Backers of startups like Town argue that an independent, model-neutral assistant can move faster and spare a business from depending on one vendor. Skeptics note that a $1 billion valuation four months after launch rests on company-reported growth, and that the big platforms can bundle similar features at little extra cost.

The signals to watch are concrete. Will Town name major customers and publish real usage numbers? Will assistants from different companies be able to talk to each other safely, or will each vendor build a walled garden? How will security teams test and approve assistants that act on their own? Will large platforms open their tools to outside agents or close them off? And as other startups chase the same idea, will customers pay for a separate assistant layer or wait for one built into the software they already use?

When Agents Meet the Outside World

At Sparknify we have been following the question of how far agents are allowed to go. In AI Agents Promised to Do Everything Online. The Web Is Locking Them Out, we examined how banks, airlines and retailers are putting up barriers to AI agents even as tech companies promise those agents will run our errands. For the most part that dispute is about agents acting for consumers on public websites, but workplace assistants will face the same test of trust within and across companies. By letting Townies email and text people outside the firm, Town has stepped right into that tension: the more useful an assistant becomes, the more places it has to be trusted.

The office of the near future may well run on conversations most of us never see, between assistants acting on our behalf.

Subscribe to Sparknify for Silicon Valley tech news and the people behind it.