AI agents are moving into your text messages. Here’s what they can do

AI assistants are starting to take requests through familiar messaging apps, then act across calendars, email and other services. A TechCrunch roundup shows how varied the pitch has become: from family scheduling to agents that can browse, run code or make reservations.

Watch Desk analysis

What happened

The services in TechCrunch’s roundup of AI agents that live in text messages include Caddy, which can turn details in messages and emails into calendar events and reminders, and Fambot, which brings family schedules and school information into app and text-message conversations. Comma, meanwhile, is described as carrying tasks across a user’s browser, computer and files, with access through messaging services including Signal, Telegram and WeChat.

The roundup also describes agents aimed at specific jobs, such as Miso for travel, alongside services that can take more open-ended actions. Folk, for example, is described as running on its own private cloud computer and handling multi-step tasks. Prices and availability vary: some products are free during beta, some offer paid plans, and Instinct remains in private beta, according to TechCrunch.

Why it matters

Messaging lowers the effort required to ask an assistant for help. But these products are not merely chat windows: many are designed to use personal information or take actions through connected services. The useful question is therefore not just whether an agent can understand a request, but what it can access, what it may do without approval, and how clearly it reports the result.

Our read

The convenience is easy to see; the boundaries are less settled. Before handing an agent access to your inbox or calendar, check what accounts it connects to, which actions require confirmation and how much the service costs once its beta welcome mat is folded away. A chat bubble is a friendly front door, not a security model.

What to watch

• Which agents move beyond beta, and what their paid plans include.
• Whether services make permissions and confirmation steps clear before taking action.
• How providers explain where personal data is processed and stored.

Discussion spark: Would you give a text-based AI agent permission to act across your email and calendar, or should it always ask before making changes?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.