Dots, Grok Bot or Muse? Count Your Jobs Before You Pick an AI Agent
Three always-on agents launched within weeks. One starts you with a single agent, one is built to run a team, and one is built for consumers. The right pick depends on how many jobs you need done at once, and who checks them.

Three always-on AI agents landed within weeks of each other. Meta launched Muse on September 8. OpenAI launched Dots at DevDay on September 29. xAI's Grok Bot, in beta, is pitched as a team of agents rather than a single assistant. For a small-firm owner the question is not which one is cleverest. It is how many jobs you need done at once, and who checks them. This comparison uses only what the companies and the press have published.
One thing up front: all three are new, and their prices and limits are changing fast. Treat everything below as a snapshot, and check each vendor's page before you buy.
What is OpenAI Dots?
Dots are OpenAI's always-on agents. In its launch post, OpenAI says each dot runs on GPT-6 Astra, has its own cloud computer, learns from feedback over time and can connect to over 4,000 apps through plugins. You can reach your dot in ChatGPT, Slack or Teams, or hop on a voice call with it. OpenAI says texting is coming soon.
The details that matter for a small firm:
- One dot for now. You start with your primary dot. The Verge reports that for now you can only create one, though OpenAI plans to allow more. OpenAI says it envisions "teams of dots" over time.
- Which plans. Dots are rolling out to ChatGPT Pro and Business Premium. Enterprise (including Edu and Healthcare) gets a beta that is off by default until an admin turns it on. According to OpenAI's Help Center, Pro access excludes the European Economic Area, Switzerland and the UK; Business Premium covers all supported ChatGPT regions.
- Pricing. Your first dot is included in Pro or Business Premium at no extra cost, with an allowance for deeper work and extended limits for the first month after launch. OpenAI says you'll later be able to add more dots and buy more speed or monthly capacity, but it hasn't published what that will cost.
- Controls. Dots come with built-in rules for when to act and when to ask. Custom Rules let you allow, require approval for, or block actions, and an auto-review step checks risky actions. OpenAI still says "always review consequential work."
- Specialist dots. OpenAI is previewing specialist dots with their own identity for roles inside organizations, starting with focused enterprise pilots.
What is Meta Muse?
Muse is Meta's personal AI agent, launched in the US on September 8. TechCrunch reports it can send emails, book travel, fill out forms and make purchases, using Link by Stripe for checkout. It runs on Meta's Muse Spark model and works on the web, on iOS and Android, and inside WhatsApp.
Muse is free to start, with a usage limit. Meta's Help Center lists two paid plans: Power at $20 a month and Maximum at $100 a month, which raise your weekly usage. Meta says Muse isn't available in all locations yet.
It is built for consumers first. And trust is the open question. Reuters reported that internal tests showed the product stalling and exposing sensitive data without authorisation, and The Verge noted a user who said Muse handed out their address to a buyer on Facebook Marketplace.
What is Grok Bot?
xAI describes Grok Bot as "your team of always-on agents." Each Bot works on a cloud computer, signs into the tools you already use, and comes back when something needs your approval. It's in beta, available with SuperGrok plans and with paid Cursor plans, including Cursor Teams, and it has its own usage, separate from those plans.
The big difference is in the word "team." The Grok Bot docs say Bots can run in parallel, message each other, share context in group chats and pass tasks between them. All of your Bots share one cloud computer, so handoffs don't need repeated setup. Bots can also claim their own email address, and every outgoing email waits for your approval.
Dots vs Grok Bot vs Muse, side by side
- Built for: Dots are aimed at work inside ChatGPT, Slack and Teams. Muse is aimed at everyday personal tasks. Grok Bot is pitched as AI teammates you give real work to.
- How many agents: Dots start with one per user, with more promised. Muse is presented as one personal agent. Grok Bot lets you run many Bots at once.
- Where it works: Dots work in ChatGPT, Slack, Teams and by voice. Muse works on the web, in its apps and in WhatsApp. Grok Bot works in its desktop and mobile apps, with its own email address if you want one.
- Price to start: Dots are included with ChatGPT Pro or Business Premium. Muse is free with limits, or $20 or $100 a month. Grok Bot comes with SuperGrok or paid Cursor plans.
- Approvals: all three have them. Dots use rules plus auto-review. Grok Bot holds drafts and outgoing email for your OK. Muse asks you to connect apps one at a time.
Why one agent versus many is the real difference
For a one-person or ten-person firm, the deciding question is how many separate jobs you want handled at once.
A common pattern for small firms that run several agents is to give each one a single narrow job: one finds leads in one market, one scores them, one researches and drafts, one tracks the pipeline, one routes incoming work. Each agent then needs only the permissions and context for its own job, and each one is easier to check. What such a team looks like, desk by desk, is set out in one giant AI assistant or twenty narrow ones.
You can't build that with a single agent. One dot can juggle several projects, and OpenAI says so. But one agent with every job ends up with every permission and every context, which is the risk a narrow-agent design avoids. Today, if you want several narrow agents that hand work to each other, Grok Bot's documentation describes that as standard. OpenAI says Dots will get there. Muse doesn't claim to.
Is ChatGPT Dots worth it for a small firm?
It can be, if your firm already lives in ChatGPT, Slack or Teams and you pay for Pro or Business Premium anyway. Then your first dot costs nothing extra, and one well-chosen job is a cheap test. It's a harder sell if you'd be upgrading just for Dots, or if you're in the UK or EEA on Pro, where it isn't offered.
Whichever agent you try, the process decides more than the product. An agent can only follow work that someone has mapped, so map the job before you pick the tool. For a practical starting point, see AI agent setup for small firms and the three agents worth setting up first.
Good first uses for any of the three
- Turn incoming email into records that a person approves.
- Score new enquiries or leads against your rules, so only the good ones get your time.
- Research and draft replies or follow-ups, saved unsent for a person to send.
- Recurring checks, like a morning summary of your calendar. OpenAI's Help Center gives this as an example of a scheduled dot task.
OpenAI's own examples for Dots include turning customer feedback into tested fixes, revising launch materials, updating research analysis, keeping a sales proposal current, and turning an interview transcript into show notes and social posts.
Frequently asked questions
What is OpenAI Dots?
Dots are always-on AI agents in ChatGPT, launched on September 29, 2026 at OpenAI DevDay. Each dot runs on GPT-6 Astra, has its own cloud computer, connects to over 4,000 apps and can be reached in ChatGPT, Slack, Teams or by voice.
How much does OpenAI Dots cost?
Your first dot is included in ChatGPT Pro or Business Premium at no extra cost, with extended limits for the first month after launch. OpenAI says more dots and more capacity will be available later but hasn't published those prices.
OpenAI Dots vs Grok Bot: what's the difference?
Dots currently give you one agent per user inside ChatGPT, Slack and Teams. Grok Bot lets you run many Bots at once that message each other, share one cloud computer and hand off tasks, according to xAI's documentation.
ChatGPT Dots vs Muse: which is better for a business?
Dots are offered on ChatGPT's paid business plans and work in Slack and Teams. Muse is a consumer agent that starts free, with $20 and $100 monthly plans, and Reuters reported internal concerns about how it handled sensitive data. For client work, a tool with business plans and admin controls is the safer starting point.
Is ChatGPT Dots worth it?
If you already pay for ChatGPT Pro or Business Premium, your first dot is a free test. Point it at one mapped, repeatable job and keep a person approving anything a client will see. If you'd be upgrading just for Dots, map the process first and see whether one agent is enough.
What are good ChatGPT Dots use cases?
OpenAI's examples include turning customer feedback into tested fixes, updating proposals as requirements change and turning interview transcripts into show notes and social posts. For a small firm, inbox-to-record work, scoring enquiries and drafting replies for review are sensible first jobs.
How do I set up ChatGPT Dots?
According to OpenAI's Help Center, you create your dot in the ChatGPT desktop app, on desktop web or in the mobile app, follow the onboarding prompts, give it a name and connect your apps. Messaging channels such as Slack are set up from desktop. Rollout is gradual, so access can take several days.
Not sure which one fits?
Start with the job, not the product. Tell Ada about one loop of work that still runs through you. She'll help you judge whether it needs one agent, several, plain automation or nothing, and put it in your free report. Talk to Ada, or see BuildBrain's services.
Sources: OpenAI, "Introducing dots" and the Help Center article "Getting started with your dot"; The Verge; TechCrunch; Reuters; Meta Help Center; xAI, "Introducing Grok Bot" and the Grok Bot docs; Cursor Help. Product details change quickly; check each vendor's page before you buy.
Similar Articles
What does AI really cost a small firm? Run the calculator →
AI automation agency for small firms →
Business process mapping examples →
Microsoft Copilot cost for a small business →
AI consulting services for small business →
Free AI readiness assessment →
Pilot purgatory: the 90-day AI proof of concept test →


