Squire. Why Now?

As LLMs have grown, early adopters have built strong habits around chatbot interfaces. Driven by consumer adoption of systems like ChatGPT, Claude (Code), and Cursor, users have built implicit understandings how context works, or how to structure and phrase prompts. At the same time, we’re still early; consumer AI is still clustered broadly around ChatGPT, and there’s space for alternatives that meets users where they’re most used to it.

Everyone texts. Rather than building another AI app competing for attention, we're inserting comprehensive AI assistance into the communication pattern people have used for years. Users text Squire like they would a human assistant: "Schedule a call with John for next Tuesday" or "Forward me that contract from last week" or “Summarize the market research report and send it to Colin” or “Move up all my meetings today after lunch by an hour”. Squire plugs into Google Calendar, Gmail, and Drive to actually execute these requests, and proactively updates users when emails come, meetings change, or it finishes work assigned to it.

We're building Squire to handle everything a human assistant can do - emails, calendar, scheduling, research, outreach. If a human can do it at a desk, Squire can do it. The difference is people can delegate by text from anywhere, and trust it gets done without having to check themselves. With Squire, we want to create the bridge from human-in-the-loop systems to truly autonomous systems by designing systems focused on allowing AI to handle known, easy-but-annoying problems that we’d be happier off not doing. We want humans to trust AI, so that one day, we’re comfortable handing them harder problems.