Loading...
Meta testing AI agent that can book restaurants and goods
Hatch is currently limited to Meta employees

Meta testing AI agent that can book restaurants and goods

Aug 28, 2026
05:05 pm

What's the story

Meta is testing a new personal AI agent, dubbed Project Hatch, to its employees for testing. The innovative tool goes beyond just answering questions and can perform a range of online tasks on behalf of the user. These tasks include booking restaurants, shopping, and even finding a dog sitter. The details about Hatch were revealed in an internal memo from Meta that was accessed by Business Insider.

Advanced functions

More than just a chatbot

Project Hatch is not your average chatbot. It comes with its own computer and can interact with websites and online services.

The AI agent can fill out forms, make purchases, and conduct detailed research.

It can also connect with services like email, calendars, Spotify, OpenTable, and Instagram.

Plus, it offers a high level of personalization where users can name their agent or specify what information to focus on.

Strategic move

Project Hatch aligns with Zuckerberg's vision of personal AI agents

Project Hatch is part of CEO Mark Zuckerberg's larger vision to create personal AI agents.

He has said that Meta wants to build "personal superintelligence for everyone."

While the company hasn't confirmed its pricing or launch date, previous reports suggested it could be a paid consumer product costing as much as $200 a month.

The project comes at a time when Meta is rethinking its internal use of AI technology.

ADVERTISEMENT

Restructuring effort

Project OT: Meta's plan to become more AI-focused

Earlier this year, Zuckerberg and other top executives devised a restructuring plan called Project OT.

The plan proposed cutting some teams by up to 60% and replacing traditional management layers with smaller groups of employees using AI tools and agents.

However, the company only implemented a 10% workforce reduction in May before canceling a second planned round of cuts later in the year due to employee backlash and doubts over whether these agents were delivering expected productivity gains.

ADVERTISEMENT