Messaging AI agents take on personal, family and work tasks

A growing group of AI agents is moving into the messaging tools people already use, including SMS, iMessage, WhatsApp, Telegram and email. These services are designed to accept requests in ordinary language, retain context, connect to calendars and other applications, and carry out actions such as scheduling, research, reservations, shopping and reminders.

The category spans general-purpose assistants and specialised products for households, travel, content work and professional workflows. Instinct is among the most highly valued entrants: after raising $350 million at a $2.5 billion valuation, it raised $1 billion in September 2026 at a $10 billion valuation. The company remains in private beta.

Messaging becomes the agent interface

For users, the premise is to avoid adding another destination app or inbox. Caddy, for example, operates in iMessage on iPhone and RCS on Android. It connects to calendars and conversations, identifies items that may need action, and can add events, set reminders, track follow-ups and conduct research. Its public beta has been available since April 2026.

Folk is reachable through iMessage, WhatsApp and Telegram, and can manage reminders, research, flights, email and restaurant reservations. The service says it runs on its own private cloud computer, can run code and can execute multi-step tasks. Its beta launched in May 2026, with a Pro tier priced at $8.33 per month for unlimited background tasks.

Instinct can be contacted by text or voice and connected to services including email, calendars and Google Workspace. The company says early users have planned trips, bought groceries, booked tickets and cancelled subscriptions. In September, it began giving users dedicated email addresses for their assistants, enabling an agent to sign up for services, contact businesses and follow up without using a personal inbox. That degree of autonomy has also raised privacy and security concerns.

Specialists target household and work coordination

Family-focused products are applying the same model to school messages, sports, meals, chores and household calendars. Fambot works through an app and text messages, sending a nightly summary of the next day’s events, to-dos and requirements. It currently connects with Gmail and Google Calendar; Outlook and Apple Calendar support is planned. The beta launched in early September 2026 after the company raised $3.5 million in pre-seed funding.

Ohai turns messages, forwarded emails and voice requests into schedules, reminders and plans, while Ollie consolidates household calendars, emails, tasks and reminders into text updates. Ollie launched in June 2026 and is one of the first mainstream family-focused AI assistants to achieve SOC 2 compliance. Orbits adds tasks such as web searches, purchases, restaurant bookings, appointment changes and requests for household-service quotes.

Other agents narrow the scope further. Miso uses iMessage for travel planning and flight arrangements, taking account of preferences and loyalty points while combining AI with a dedicated travel team. Town is aimed at professional work, offering a “Townie” that learns recurring workflows across email, calendars, documents and Slack. The company raised a $55 million Series A in June 2026 led by Andreessen Horowitz and Forerunner.

Autonomy requires operational boundaries

The technical distinction is increasingly not the chat response but the ability to connect services and complete multi-step work. Wajo’s Fo has its own email address, phone number and payment card for interacting with businesses, and Wajo says a human assistant can be brought in when the agent cannot complete a task. Poke, meanwhile, became the first AI agent approved for Apple Messages for Business in June 2026 before parent company The Interaction Company of California was acquired by Cognition a month later.

For businesses, the practical question is which data sources an agent can access, which actions it may take and how permissions, identity and escalation are controlled. Messaging-based agents may simplify routine coordination, but delegating work involving customers, purchases or accounts requires clear boundaries before automation is put into daily use.