Town raises $108M: AI assistants for teams need sharing rules

Town, the San Francisco startup behind the "Townie" AI assistant, raised a $108 million Series B at a $1 billion valuation, co-led by Forerunner Ventures and Menlo Ventures. The pitch has changed. A Townie used to work for one person. Now Townies work across a team: they ask each other questions, hand off tasks, and email people outside the company. AI assistants for teams are an org chart for software. Somebody has to decide what each one is allowed to say.

What actually happened

Per The Next Web and Town's own announcement:

• The round: $108M Series B, announced October 19. a16z and First Round Capital came back in. Town raised a $55M Series A at its public launch in June.
• The product: each person gets a Townie that connects to email, calendar, documents, and core work systems, and learns their contacts, projects, and writing style.
• The team feature: Townies can draft in an executive's voice by working with that executive's Townie, find which teammate knows a prospect and ask for an intro, and test a draft against teammates' Townies before a meeting. When one person improves a process, every Townie on the team can pick it up.
• Outside contact: Townies can talk to customers, partners, and candidates by email and text.
• The guardrails Town describes: a Townie shares only what its owner would share, asks permission when needed, and sends decisions back to its owner. Town says it does not train models on customer data by default, picks a model per task from several labs, and is SOC 2 Type 2 compliant with a public DPA.
• Traction: seven-figure ARR within weeks of launch, and since July, 4x growth in paying customers and 9x in paying teams, per the company. Town did not publish pricing in either announcement.

Why AI assistants for teams matter for your business

"Shares what its owner would share" is a judgment call. Your bookkeeper's assistant knows payroll. Your sales lead's assistant knows margins. When one agent can ask another agent a question, the permission model is the product. Before you turn this on, write down what each role can tell each other role. Town's public page does not spell out the specific permission controls, so ask for them in the demo.

Agents that email customers speak for you. A Townie that sends a quote follow-up or a candidate reply is your brand in the inbox. Decide which messages go out on their own and which wait for a human.

The good part is portability. Town says customers own the context their Townies build and can take it with them. That is the right default. Test it: ask for an export and see what you actually get.

Shared process updates cut both ways. One teammate improves the call-prep checklist and everyone gets it. One teammate adds a bad step and everyone gets that too. Keep the playbook in a document someone owns.

Sources: The Next Web, Town.