Meta Muse Voice Calls vs Grok Bot Phone Calls: How They Work

How do Meta Muse voice calls and Grok Bot phone calls work?

They solve different problems. Meta Muse now lets you talk to the assistant by voice or video call from inside the app, using a call button in the chat. The assistant can draw on connected accounts such as Gmail and calendar. Grok Bot goes the other direction: you ask it to phone a number, and it makes the call for you, for example to ask a business about its hours.

What can Meta Muse voice and video calls do?

Meta Muse added two things: voice chats and video calls. In the demo, the user opened a new side chat and found a call button in the top right corner. Pressing it started a live conversation with a voice assistant.

The assistant greeted the user by name and noted the day of the week. It also took spoken instructions about itself. The user said the voice name didn't match the voice, and the assistant replied that it had switched to a mid-pitched, warm, middle-aged male voice and asked how the new one sounded. That is a small detail, but it shows the call mode accepts changes to its own persona mid-conversation.

The more important part is what it can reach. Asked what it could help with, the assistant listed:

• Gmail and calendar handling
• Content ideas for an Instagram account
• Research
• Drafts
• Scheduling

It then offered to check the inbox, the calendar, or pull recent content ideas. It also summarized items from today's refresh, including a same day hands-on with a new Gemini video model.

The demo's creator described the voice mode as giving more of a "talking to an assistant" feel than typing. He also noted that Muse can do everything it could already do, just by speech. He did not read out his emails or calendar on camera because the answers could expose sensitive information.

Does the Muse avatar matter?

Calls display a small avatar while you talk. The creator was blunt that he didn't see the point of it. Based on that demo, the avatar looks like a cosmetic layer on top of the voice conversation, not a separate capability. Whether a video call adds anything beyond voice (such as the assistant seeing you or your camera feed) was not shown, so treat that as unverified.

How does Grok Bot make phone calls for you?

Grok Bot recently rolled out a feature that lets it place phone calls for the user. When the creator asked whether it could make calls now, it answered that it could.

His test was simple. He gave it a phone number (his own, blurred in the video), told it to treat the number as a business, and asked it to find out the business's hours that day. The idea was to hear what the bot sounds like from the receiving end. The transcript cuts off right as the test begins, so the source does not show the call itself, how natural the voice was, whether it handled follow-up questions, or what it reported back.

What we can say from the source:

• The feature exists and Grok Bot itself says it can make calls.
• The obvious use case is a short, factual errand such as checking opening hours.
• The creator had not tested it before this run and said he was a fan of Grok Bot as an agent in general.

What we can't say: pricing, which countries or numbers are supported, whether it discloses itself as an AI, how it handles voicemail or phone trees, or whether it asks for approval before dialing. Check the product's own documentation before relying on any of those.

How do the two features compare?

Which one is more useful?

It depends on the job.

Muse's call mode is useful if you already live in Gmail and a calendar and want to ask "what should I do next?" while your hands are busy. It replaces typing with talking. It doesn't add new abilities, so if you were happy with the text version, the value is convenience.

Grok Bot's phone calls address a task that text assistants have never handled: reaching a person or business that only answers the phone. Confirming hours, asking about availability, or checking a price are low-stakes and repetitive, and a bot is a reasonable fit. That is a real capability gap being closed, which arguably makes it the more novel of the two. The caveat is that the evidence here is thin. Until someone shows the full call and its summary, the usefulness is a promise, not a result.

What are the limits and risks?

Both features put an AI into a more personal channel, and both raise practical questions.

• Privacy with Muse. Connected email and calendar mean a spoken question can pull sensitive details aloud. The creator avoided asking for that on camera for this reason. Use call mode where others can't overhear.
• Accuracy by voice. Speech hides the sources and the text that you would normally skim. For anything consequential, ask for a written summary afterward.
• Third-party consent with Grok Bot. A bot calling a business raises disclosure and recording questions that vary by location. The source doesn't say how Grok Bot handles them.
• Errors without a human in the loop. If a bot mishears a time or a price, you may not catch it unless it reports back clearly.
• Scope. Neither demo showed a complex, multi-step task completed from start to finish.