How it works

Everyone on your team gets their own agent. You teach it the calls you have, it joins them, and it gets better after each one.

Before the call

• Write a playbook by talking. Describe a kind of call in your own words: the goal, what to find out, when it's done. It turns that into a playbook you can edit.
• Attach what matters. Add documents and links to a playbook, like a contract, a price sheet or the other side's redline. It learns them for those calls only, including where each side stands.
• Give it your material. Files, notes and your website teach it about your product. You can correct anything it learned.

On the call

• It joins Google Meet, Zoom and Microsoft Teams. Paste a meeting link, or connect your calendar and it joins your calls on its own.
• Run the call, or chip in. It can lead a conversation like a discovery call or a customer check-in, or stay quiet and answer when someone says its name.
• Show, don't just tell. It can share its screen to show your product live, and draw a quick sketch when someone asks for one.
• Always visible. It joins as a participant with (AI) after its name, as a calm dot pattern or a talking face.

After the call

• The transcript. Every line, kept in your workspace.
• What it found out. The answers to each thing your playbook asked for.
• What it learned. Lessons for next time, which you can edit or remove.

Your team

• One agent each. Everyone names their own agent and picks its voice.
• One shared memory. What one agent learns about your product, every agent knows. A correction reaches all of them at once.
• Spending you can see. Each workspace has a monthly limit, and a usage page shows what every service costs.