Modern Creator
Paul J Lipsky · YouTube

Gemini's New Personal Agent Is Here (Muse Killer)

Google's new Gemini agent, called Task, takes work by text and email. It's capable, but its boring name and chat-first design keep it from feeling like a real assistant.

Posted
3 days ago
Duration
Format
Review
educational
Views
238.8K
2.9K likes
Big Idea

The argument in one line.

Google's Gemini Task agent matches Meta Muse on capability, but its generic name and feature-style design only start to feel like a personal assistant when you use it through text messages.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You live in Google's ecosystem (Gmail, Calendar, Slides, Sheets) and want an agent that works with those tools out of the box.
  • You're weighing Gemini's new agent against Meta Muse or similar personal agents and want a direct comparison before committing.
  • You want the exact phone settings steps to turn on text and email access to Gemini.
SKIP IF…
  • You're in the UK or European Economic Area without a Pro or Ultra subscription, where the free tier isn't offered.
  • You want a deep technical look at the model or agent architecture; this is a feature walkthrough and opinion review.
TL;DR

The full version, fast.

Google released a personal AI agent inside Gemini, named Task, that competes with Meta Muse. You can text it, email it at gemini at google.com, or CC it, and it works with Google Slides, Sheets, Calendar and connected apps like Instacart. It's proactive, can delegate to other tasks, and is free in the US, Canada, Australia, India, Japan and South Korea. The reviewer's complaint is that Task has no name, avatar or persistent chat in the app, so it feels like a feature. Texting it fixes much of that. Occasional silent non-replies remain a flaw, and Muse still has more third-party integrations.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →
Chapters

Where the time goes.

00:00 – 00:31

01 · Intro

Google launches a personal agent to rival Muse; the reviewer teases the odd design choices.

00:31 – 01:06

02 · Who gets access

Free in six countries; Pro or Ultra elsewhere; UK and EEA excluded.

01:06 – 04:07

03 · Texting and tasks

Text the agent, hand it work, use it from CarPlay, and watch it build Google Slides and act proactively.

04:07 – 05:30

04 · Email and payments

Email or CC gemini at google.com; Google Pay purchases coming.

05:30 – 06:46

05 · Sponsor: Semrush

Semrush AI Visibility toolkit for tracking brand mentions in AI answers.

06:46 – 09:19

06 · The big issue

The name Task, no avatar or persona, and tasks mixed with regular chats.

09:19 – 11:26

07 · Set up Gemini

Gemini app settings: Channels, text and email with Gemini, personal intelligence, connected apps.

11:26 – 13:46

08 · More features and limits

Image generation, dashboards, orchestration, cross-app handoff, and the silent non-reply bug.

13:46 – 15:04

09 · Gemini vs Muse

Not a Muse killer yet, but a strong start; Muse has more integrations and focus.

Atomic Insights

Lines worth screenshotting.

  • Google's new Gemini agent is free for all users in the US, Canada, Australia, India, Japan and South Korea, with no Pro or Ultra subscription needed.
  • Texting the agent like a friend is the best way to use it, because you get one long continuous conversation instead of a new task every time.
  • You can CC or BCC gemini at google.com on an email, tell the other person your assistant will handle it, and the agent knows what to do.
  • The agent is an orchestrator: for bigger jobs it opens a new task and delegates the work rather than doing everything itself.
  • A task started in the Gemini app can be picked up from text messages, but only if you started it as a task and not a regular chat.
  • Merging Gemini Spark into the main app under the generic name Task traded a distinct assistant identity for integration.
  • The agent is proactive: 45 minutes after failing to make a call, it came back with a website check and offered to draft an email.
  • The biggest reliability bug is that it sometimes acknowledges a task, never replies, and has already finished the work.
  • Phone calls and Google Pay purchases are announced but not yet enabled, so booking by phone isn't available at launch.
  • Connected apps are fewer than rivals offer (Instacart, OpenTable, Peloton, SeatGeek, Shopify), which is where Muse still leads.
Takeaway

A capable agent held back by its packaging

WHAT TO LEARN

Gemini's Task agent is feature-competitive with Muse, but whether it feels like an assistant depends on using it by text, not inside the Gemini app.

  • Check availability first: it's free in the US, Canada, Australia, India, Japan and South Korea, needs Pro or Ultra elsewhere, and excludes the UK and EEA.
  • Hand it work instead of asking for advice, for example triage the inbox and draft replies, or plan dinner and fill an Instacart cart.
  • Use text messaging as the main channel, because it gives one continuous conversation and more proactive follow-ups than the in-app version.
  • Forward or CC gemini at google.com on emails to delegate scheduling, report-finding and tracking without leaving your inbox.
  • Set it up from the Gemini mobile app under Settings, Channels, then turn on memory in personal intelligence and connect as many apps as you want.
  • Start work as a task, not a chat, if you want to resume it later from text messages.
  • Expect reliability gaps: slow tasks may finish without a notification, and phone calls and Google Pay aren't live yet.
  • Judge agents on identity as well as features: a name, persona and single running thread are why Muse feels like an assistant.
Glossary

Terms worth knowing.

Gemini Task
Google's new personal agent inside Gemini. It takes work you hand it and carries it out across Google apps and connected services.
Muse
Meta's personal AI agent, which has a name, avatar and personality and keeps one long chat with the user.
Orchestrator
An agent that splits a job up and hands pieces to other tasks instead of doing all the work itself.
Resources

Things they pointed at.

01:43productGemini Spark
02:00productMeta Muse
11:00productInstacart, OpenTable, Peloton, SeatGeek, Shopify (Gemini connected apps)
Quotables

Lines you could clip.

06:40
“You can't give it a name, a personality, or an avatar.”
Crisp critique of the product's biggest design gap→ TikTok hook↗ Tweet quote
08:41
“Using the text message feature completely transforms this experience from something that feels more like a feature into feeling more like the type of agent we're used to with Muse.”
The key tip in one line→ newsletter pull-quote↗ Tweet quote
The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

Google finally did it. They released their own personal AI agent to compete with tools like Muse and maybe even Grokbot. I've been testing this out for the past couple of days and I absolutely love it.
However, as you'll see, some of the choices they made as to its features and the way that it works are a little bit interesting. And I think those choices are either brilliant or it's going to end up hurting them. You'll see what I mean by that when I break this all down for you.
So let's go ahead and do that. I'll show you what the new agent looks like. how it works, and what you can actually do with it.
Let me start out by telling you who will have access to this. And I have some good news for most people, but not everyone. Google is making this new agent available for free to all users in the United States, Canada, Australia, India, Japan, and South Korea.
So if you're in one of those countries, you don't need a pro or ultra subscription, you can use it for free. It is rolling out though, so you may not see it in your account immediately. Outside of those countries, Google says you will need a pro or ultra subscription, and there are some excluded regions, including, big surprise, the UK and the European Economic Area.
Okay, now let's actually take a look at this agent together. There are a couple of different ways you can interact with it, but my favorite way right now is to text message it. You can message it the same way you message a friend using the messaging app on your phone.
So this works for both iOS and Android. And as you would expect from a personal AI agent, this is an agent that you give work to. So instead of asking it for advice on how to manage your email inbox, you tell it to triage your inbox for you.
It will go through everything and draft the replies. Instead of asking for help with planning out your dinner, you can tell it to plan your dinner and then to add all the ingredients to your Instacart for you.
None of that is really new though. You could already do that with Gemini using tools like Gemini Spark, but this new agent has new capabilities that make it more similar to something like Muse, like being able to text it, for instance, which honestly I think is a lot more convenient and it makes it very easy to send tasks off just by telling the Apple Assistant to message Gemini.
And it's going to do that because it can send text messages to anyone. So you say, hey, send a text message to Gemini and tell it to draft an email for someone or to start putting together a new meal prep for the week and start adding ingredients to my Instacart. And because you can use your assistant for that, you can even do it while you're driving using something like CarPlay.
Here, for example, I sent it a text message and asked it how people are feeling about the new model that recently came out from OpenAI. After it gave me that breakdown, I told it to put together a presentation for me. And then here you can see that we have that presentation.
And what's great about this new agent is that it already works with your existing Google ecosystem. So Google already has Google Slides. It already has Google Sheets.
So it's going to use those tools. Here, it created for me a presentation inside of Google Slides. And all of that works out of the box.
Just like with Muse, phone call support is coming to the Gemini agent. So it's gonna be able to do things like book appointments for you over the phone, and then add the date and time directly to your Google Calendar. I tried this out.
I told it to call my local climbing gym to ask if they're gonna be open on a holiday. And it said that unfortunately I don't have phone calls enabled in my account yet, but it should be coming soon. This agent is also proactive.
So after I asked it to make that phone call and it said it couldn't for me, it then about 45 minutes later sent me this message. And it said it looked on the gym's website and it found that there was no calendar listed with the holiday hours. But it did suggest that it could draft a quick email that it could send to the climbing gym to ask them if they're going to be open on Thanksgiving.
And then it asked for my permission to do that. And again, it did that completely on its own about 45 minutes after it told me that it couldn't make the phone call. You can also email your agent or CC it in emails.
So let's say you receive an email and it's asking for your availability. What you can do is click on forward. and then type in gemini at google .com.
And because it's coming from your email address, Gemini will know that it's coming from you. And then you can give it the task. So I might say, find some availability on my calendar and then email me back with the time.
Or it might be something like, they're looking for that report, find it and send it to them. You know, different things like that. Or you can even reply to the email and CC them in it.
So I could see this being really cool. reply at a cc or bcc at gemini at google .com and then you might say to the person that sounds great i'll have my assistant put together the report and send it to you and then your agent's going to know that that's what you want them to do after you send an email your agent will email you back here you can see an example of that I told it to keep track of something in an email and it responded by saying, I'm working on this for you.
I'll track the updates and you can see the completed task in Gemini. And finally, you'll be able to set up payments through Google Pay, so your agent will be able to make purchases for you with your permission. So, so far, there's a lot to love about this new agent, but there are a couple of things that I don't love.
Before we talk about that, though, let's quickly hear from today's sponsor. Right now, a lot of people are asking ChatGPT or Gemini what to buy or who to hire instead of searching Google. And the AI usually just names a few options.
Ask ChatGPT for the best project management app, for instance, and it names just a handful. So if your business isn't one of those, you're not even in the conversation, even if you rank well on Google. So if you're a business owner, this is something you have to be paying attention to.
And that brings me to the sponsor of today's video, SEMrush. SEMrush is a platform that shows you how visible your business is in search. And its AI visibility toolkit shows you how often ChatGPT, Gemini, and Google's AI answers are mentioning your business and where they're mentioning your competitors instead.
So let's say this is my business. I put in the website and get an AI visibility score from 0 to 100. That's how much my brand is showing up in those AI answers.
Then in competitor research, I can see the questions where my competitors show up and I don't. Those are the gaps I'd go after first. Now, SEMrush can't guarantee an AI will recommend you, but you'll know where you stand and what to work on.
You can try this with a seven -day free trial of SEMrush One, which includes the AI Visibility Toolkit. Just use the link in the description. Now let's get back to the video.
Okay, so far, so good. Just looking at the features alone, this sounds like a very capable agent that could potentially rival Metamuse. But what does it actually look like when you're not texting it?
And what is it called? Well, this is where I think Google may have made a mistake. Google's new personal agent can be found here inside of Gemini.
Do you see it yet? It's right here. It's called Task.
That's it. You can't give it a name, a personality, or an avatar. And there's no separate section for it either.
So all of your regular chats and your tasks live together intermingled. And I don't know how I feel about this. It's kind of boring.
Part of the reason Muse works so well is because It feels like an actual personal assistant. You give it a name, an avatar, it has a personality, and you can have one long chat with it.
From a branding perspective, the name Task is, I think, an interesting choice. So many times I've said, oh, I'll just have Muse do that for me, or I'll have one of my bots do that. But Task, I don't know what I'm gonna say.
Will I say I'll have Task do it for me? I don't know, I'm not totally sold on it. And then because of the way it's set up, you're gonna be starting a new task.
every time you use it, which is a lot different from Muse, where primarily you just have one long single place where you talk with it. So Task really doesn't feel like an assistant as much as other tools like Muse, Grokbot, Duet, or Dots does. What's particularly interesting about this is that there actually was a better name for this in the past.
Task used to be Gemini Spark, and it was sort of separate from the main Gemini. And now they've merged it together and changed the name. So I don't quite love that.
I would like it to feel more like a personal assistant that's sort of separate from Gemini in a way, but there is a fix for this. Using the text message feature completely transforms this experience from something that feels more like a feature into feeling more like the type of agent we're used to with Muse. And I highly recommend if you wanna use task to use it from your text messaging app because it is a much better experience.
If you look here, I have one long continuous conversation with Gemini in here. And this is also where I've noticed it is more proactive than just using it natively inside of Gemini. So using it this way feels much more like using Muse.
So what I think I'll do now is I will show you how to set up both the text messaging feature and a lot of these other features so you can get the most out of the new task. So to set this up, you need to use the Gemini app on your phone. Unfortunately, you cannot set this up on the computer, at least not as of right now.
So go ahead and open up the Gemini app and then open up the side menu and you want to come into the settings by clicking on the gear icon. Now scroll down and you should see a new preference option right here called channels. Clicking on this, you will see two options, but up here at the top, you will see text with Gemini.
I've already set mine up, but you'll see a button right here to set yours up. Click on that and you can enter in your phone number if it's not in there already. It will then send you a verification number that you will put in.
After that, you'll want to click on this button that says send text. That will start the conversation. inside your messaging app.
One tip for you, mine went to the spam folder. So if you don't see it, check your spam folder and then mark it as a known number or add it to your contacts so that it doesn't end up in spam again. And then that's it.
After that, it should work and you'll be able to message it the same way I've been showing you already. Underneath that, you will also see the option to email with Gemini and all you have to do is turn that on. And then you can send it tasks just by emailing it, like I showed you before, by sending emails to gemini at google .com.
While we're in the settings, let me show you a couple of other things. So I'll back out of this one. You'll also wanna come into personal intelligence and make sure you have as much of this turned on as you want.
So if you turn on memory, that will allow Gemini to learn more about you as time goes on, and that will help it become more helpful for you. You also wanna come into connected apps. And the more apps you have connected here, the more it's gonna be able to do for you.
So if we come over to all of these, you can see what is currently available. There are a few in here, not as much as other tools, but there are adding more of them as time goes on as well. You can see things like Instacart, OpenTable, Peloton, SeatGeek, and Shopify, for instance.
Now you're ready to start sending it text messages. Now I already showed you a bunch of things that this can do, but let me show you a couple of other cool things that I've discovered. So first you can have it create images for you directly inside your text messaging app.
Videos, however, don't appear to be supported at least as of right now. I also showed you before how it can create Google Slides for you, but then I asked it to put together a dashboard for me. You see, I'm thinking about buying a new pair of AirPods.
I wanted the best deal on it. So I wanted a dashboard to track all the prices that would be updated for me every morning. Here's the response, and this revealed something else really cool that this agent can do.
If I click on open for this dashboard, you'll see that it opens up a task or chat directly inside the Gemini app. And of course, if I open up this, that actually shows me the dashboard that was created. The dashboard's cool and all, but what I think is even cooler is that this reveals that the agent is actually an orchestrator.
Some of the tasks it's just gonna do on its own, but for others, it will open up a new task and tell it to do the work. In other words, it will delegate the work out to other tasks. Also, if you start a task from within the Gemini app, you can pick it up from within your text messages.
So for instance, right here, I came into the Gemini app and I started a new task and I told it to write a scary story for me. Then I went over to the conversation here in my text messages, and I told it to look for that scary story and to send me the first paragraph of it. And here it is.
It was able to find that for me. For that to work, though, you would need to start it as a task in Gemini and not as a regular chat. You'll also notice here, though, that it's not perfect.
The biggest problem I've seen with it is that sometimes I will send it a task, it will react to it to let me know that it saw it, but then it will never respond. and I'll have to ask it for an update before it lets me know that, oh, actually it already finished that work. It just never told me.
And that happened quite a few times. So I do hope they fix that. So it actually does respond once it's finished.
It seems to happen a lot when the task I give it takes a little bit longer than something that's faster. So is this a meta Muse killer? Well, short answer, no, but I think it's a very strong start.
It has a lot of similar features as Muse and actually the model that powers Gemini task. is more powerful than the model that powers Muse. And both of them are also free.
However, Muse works with a lot more third -party tools than Task does. And also it's focused on being a personal agent as opposed to more of a feature, I think is one of its greatest strengths over Task. But Task is off to a great start.
And if you're already in the Google ecosystem or you don't trust Meta and you do trust Google, then this could be the agent you've been waiting for. So that is the new Gemini Personal Agent, a fantastic tool with a rather boring name. But who knows, maybe they'll call it Spark again and create a standalone app for it and really transform it into something much more like Muse.
We'll see. Otherwise, I'm going to continue to use it through my iMessaging app because I do think it is a great tool. Feel free to share your thoughts in the comment section down below though.
And then if you wanna learn more about what you can do with MetaMuse and maybe some of the capabilities that it has that you cannot do with Gemini Task, check out this video right here. This is a video I made recently going over some of my favorite Metamuse use cases. So definitely check this out and I'll see you over there in one second.
Bye for now.
The Hook

The bait, then the rug-pull.

Google just released its own personal AI agent, and after a few days of testing the reviewer loves it. But two design choices, the name and where it lives, may either be brilliant or hurt Google.

CTA Breakdown

How they asked for the click.

VERBAL ASK
14:58next-video
“Check out this video right here on my favorite Metamuse use cases.”

Soft end-screen push to a related Muse video after asking for comments.

MENTIONED ON CAMERA
Storyboard

Visual structure at a glance.

open
hookopen00:00
texting the agent
valuetexting the agent01:06
Task name issue
valueTask name issue07:02
setup in settings
valuesetup in settings09:42
verdict
ctaverdict14:00
Frame Gallery

Visual moments.

One-click upgrade to your Google

Get more breakdowns in your search results

Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.

Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
Watch next

More from this channel + related breakdowns.