Muse gets AI agent UX right
Claire reviews Meta's Muse personal AI agent, praising its consumer-friendly design, intuitive UX patterns, and thoughtful features like identity customization, activity feeds, and goal tracking. While impressed by the overall experience and artifact generation quality, she identifies browser-based shopping as an area where Muse underperforms compared to competitors like Claude.
Summary
Claire provides a detailed first-pass review of Muse, Meta's new personal AI agent positioned for consumers—particularly parents managing busy household schedules. She walks through the onboarding experience, which begins with seamless Facebook account integration and immediately establishes trust through permission-granting practices. Key features highlighted include the ability to name and customize an agent's identity with editable soul and memory parameters, an activity feed that tracks all tasks performed with full lineage and tool usage visibility, and a goals system that lets users set personal objectives across health, relationships, finance, and productivity categories.
Claire demonstrates Muse's capabilities through several use cases: setting up calendar management for her children's activities, requesting a personalized PDF morning newsletter that combines family schedules with local news and conversation starters, and exploring the podcast generation feature (similar to NotebookLM). She creates a custom agent named Slime with a teal dragon avatar and tests shopping functionality for IMAX movie tickets, which succeeds, and shoe shopping, which struggles with accuracy.
The reviewer emphasizes that Muse's strength lies in thoughtful product design rather than raw AI capability. She praises the gentle, human-first tone of the agent, the progressive disclosure of complexity (exposing browser use rather than raw VM details), and design affordances like animated avatars that change appearance based on task type. She contrasts Muse favorably with competitors like Claude/Codeex and OpenClaw, noting that while Codeex may have superior browser-use capabilities, Muse better understands consumer needs and presents information in less intimidating ways.
Claire identifies browser-based shopping as Muse's primary weakness, where it struggles to find correct product variants and retrieve accurate information. She concludes by positioning Muse as ideal for personal wellness, family logistics, and home management tasks rather than professional/chief-of-staff work, and encourages product designers and managers to study Muse's design patterns—particularly the introduction of primitives like Tasks, Ideas, Goals, and Library—as a template for consumer-facing AI agent design.
About this episode
I spent a few hours putting Meta’s Muse, its new personal AI agent, through a real first-pass test: onboarding, calendar management, goal setting, a one-shot family morning newsletter, browser-based shopping, and the animated avatar that honestly surprised me. *What you’ll learn:* 1. Why Muse is the best-designed personal agent I’ve tested, and what specifically made it feel that way 2. The one-shot family PDF Muse produced that Claude and Codex never quite nailed 3. How Muse’s permission model works, and why it’s different from every other agent I’ve used 4. Why I set up a sleep training goal in Muse, and what it revealed about agent tone 5. The activity feed feature I immediately wished Codex and Claude Code had 6. Where Muse failed, and what it says about the limits of this category right now 7. The animated avatar decision that showed me what top-of-craft AI product design actually looks like *Brought to you by:* Optimizely—Your AI agent orchestration platform for marketing and digital teams: https://www.optimizely.com/howIAI OpenArt—An all-in-one AI creation platform for images, videos, music, audio, and more: https://openart.ai/suite/chat?utm_source=online&utm_medium=influencer&utm_campaign=infl-howiai-ga-na-acq-web *In this episode, we cover:* (00:00) What Muse is and who it’s actually built for (04:41) Signing in and the onboarding flow (07:16) The activity feed and its task lineage (08:24) First real task: managing the family calendar and deleting soccer practice (09:48) Requesting a morning newsletter PDF (14:29) The personalized news feed and how I set it up (16:05) The “Ideas” feature as an out-of-the-box prompt library (17:10) Setting up personal goals (water, shoes, and sleep training) (21:40) Library: documents, websites, images, videos, and podcasts (23:11) Quick recap and what I love (23:56) Activity feed design deep dive: tool calls and step-by-step lineage (25:18) How Muse handles permissions (26:12) The animated avatar: Polly becomes Slime, the teal dragon (29:34) Browser use test: shopping for New Balance 9060s (not great) (31:15) Browser use test 2: buying IMAX tickets for The Odyssey (much better) (33:34) TL;DR and what I’ll actually use Muse for going forward *Tools referenced:* • Muse: https://muse.ai/ • Stripe Link (payment method featured in Muse): https://link.com • 1Password (future Muse integration mentioned): https://1password.com • OpenClaw (Claire’s previous personal agent setup): https://openclaw.ai/ • Grok Bot (Grok-based agent from prior stack): https://x.ai/news/introducing-grok-bot • Codex (OpenAI coding agent, comparison point): https://openai.com/codex • NotebookLM (Google, comparison to Muse’s podcast generation): https://notebooklm.google.com *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email [email protected]._
Key Insights
- Muse is deliberately positioned as a consumer product for parents and millennials with children, not young people or professional users, as evidenced by landing page use cases focused on permission slips, stroller shopping, and family schedule management
- Meta's inclusion of technical VM specifications (persistent isolated Linux virtual machines with browser, CPU, and memory details) on the landing page contradicts the consumer-friendly messaging and will confuse mainstream users who don't care about infrastructure details
- Muse demonstrates superior artifact design quality compared to previous agents like Codeex and OpenClaw, with the morning newsletter PDF featuring beautiful layouts, intelligent scheduling conflict detection, and age-appropriate news formatting that actually works as a family communication tool
- The activity feed showing every tool call, script, and step-by-step task execution serves dual purposes: transparency for consumers and developer visibility into agent reasoning, presented in an elegant UI that other agents lack
- Muse's permission-granting model uses contextual, multi-step approval rather than binary options like 'auto-approve' or 'ask every time,' exemplified by asking separately to access email, then again to use discovered information, which maintains trust without becoming annoying
- The avatar design employs image generation to create dynamic, task-aware visual feedback—where the avatar pulls up a laptop when working and generates orb-like effects during image/media generation—demonstrating design craft that goes beyond interface aesthetics into interactive micro-interactions
- Muse's podcast generation capability (creating multi-voice news briefings) was an unexpected feature that impressed the reviewer despite not being promoted, suggesting Meta's agents have broader content creation capabilities than advertised
- Browser-based shopping functionality in Muse succeeded for movie tickets but failed for shoe shopping due to inability to find correct color variants and retrieve accurate product information, indicating the vision and web scraping components have uneven performance
Topics
Transcript
[0:00] It's the weekend. The kids are out of the house, and that means one thing. It's time to test a new AI agent. Today, I'm going to give you my first pass opinion about Muse, Meta's new personal agent. We're going to talk about what Muse is, what I had it do for me, and what it didn't do quite well, and why I think it might be my favorite designed agent I played with in a long time. Let's get to it. Optimizely agent platform for marketers is built on a bigger idea. A whole directory of agents now joined by a newly revealed squad of [0:34] virtual teammates. Each with a defined role and a personality of…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
Muse made me a 6-min AI podcast
The speaker demonstrates Muse, an AI tool that generates podcasts, by having it create a 6-minute podcast about AI news from a weekend in September 2026. The tool quickly produced a conversational two-host format discussing contrasting perspectives on AI safety concerns, and impressively offers direct publishing options to Spotify and Apple Podcasts.
Grok Bot sends this design into Figma from my phone
A designer demonstrates how an AI bot can convert voice descriptions and screenshots into interactive Figma prototypes, generating multiple design iterations from natural language input. The tool interprets conceptual descriptions and produces working prototypes without requiring manual prototyping work in Figma.
Muse beat OpenClaw and Codex on design
The speaker demonstrates Muse, Meta's AI agent, successfully generating a one-page PDF newsletter for family schedules and news that surpasses designs from competing AI tools like OpenClaw and Codex. After connecting to email and allowing the AI to ingest personal family information, Muse produced superior design and content quality for the requested morning brief.
Grok Bot turns our brand book into Figma cards
A designer demonstrates how they used Grok Bot to automate the creation of marketing material cards in Figma by applying a brand book template across multiple assets. The discussion highlights how AI can eliminate repetitive production design work, freeing designers to focus on creative strategy rather than manual execution.
Google Places + Grok Bot auto-updates this website
A designer created an automated workflow using Grok Bot and Google Places API to effortlessly update their personal website with high-quality, processed images of locations. The system identifies places from photos, generates 3D miniature visual styles in light and dark versions, and automatically publishes them to the site, solving the problem of website staleness.