TechRadar writer says Meta’s Muse made a personal AI agent useful

TechRadar author Lance Ulanoff says Meta’s Muse AI agent was more useful than he expected after he connected it to his personal Gmail account. He named its avatar Charlie, though the article says the avatar is officially called Jolly.

Ulanoff describes Muse as a proactive assistant that runs in the background. He says Meta’s system uses a virtualized cloud environment, stores credentials in a “Secure Credential Store” and applies oversight intended to prevent risky activity. Those are Meta’s stated controls; Ulanoff says he felt some trepidation about granting access.

When asked to check for urgent messages, Muse found three worth reviewing: a Google security alert related to the new connection, a notice from his mobile provider about a recent account change, and an email about an embargoed communications product. Ulanoff asked Muse to summarize the last message and draft a reply. After he corrected the draft to remove a proposed call or meeting, Muse placed it in Gmail drafts. He then asked it to send the email with his work address copied.

Later, Muse found details in an email about a speaking engagement even though Ulanoff misspelled his contact’s name. He says it returned the location, parking information, a phone number and a map. The account is a single user’s early experience, not an independent assessment of Muse’s privacy safeguards.