The next AI race may not be about who gives the best answer. It may be about who can actually get the work done.
OpenAI is stepping deeper into the world of AI agents with Dots, a new product designed to work across apps, follow tasks over time, and take action on a user’s behalf.
The launch comes at an interesting moment for the AI industry.
Meta’s Muse has quickly gained attention with consumers, while Google, Anthropic and other players are pushing their own agentic AI products. OpenAI is now making its own move, but with a clear focus on professional and enterprise use cases.
That distinction could matter.
From answering questions to doing the work
Generative AI made it easy to ask a chatbot a question and get an answer.
Agentic AI is trying to remove the next step.
Instead of simply telling you what to do, an AI agent can potentially:
- Monitor information across different apps
- Identify tasks that need attention
- Take action on your behalf
- Follow a workflow over time
- Connect information from multiple tools
- Work with other agents on more complicated jobs
That is the idea behind Dots.
OpenAI says Dots can connect to 4,000 apps through ChatGPT. A user could ask a Dot to keep track of a project, add something to their schedule or follow up on an ongoing task.
In one example shared by OpenAI, a Dot noticed a software bug alert arriving through Slack and began investigating it.
In another, a Dot identified that an invoice had not been submitted, prepared it and submitted it after receiving the user’s approval.
That is a very different proposition from simply chatting with an AI.
The goal is to turn AI from a tool you operate into something that can operate alongside you.
OpenAI is taking aim at Meta’s Muse
The timing of the launch is hard to ignore.
Meta launched Muse earlier this month and quickly attracted users. The agent climbed to the top of Apple’s App Store and Google Play, while third-party app analytics cited in the source material put its early downloads in the millions.
Muse has given Meta a strong early position in the consumer AI agent market.
OpenAI is approaching the market differently.
Meta is leaning heavily into consumers. OpenAI is positioning Dots toward businesses and professional users.
Dots will initially roll out to Pro, Business Premium and Enterprise users in available markets.
That creates a fascinating split in the AI race.
Meta is showing that AI agents can become consumer products.
OpenAI is betting that they can become workplace infrastructure.
And the bigger opportunity may ultimately be whichever company can make agents useful enough that people stop thinking about them as “AI products” and start thinking about them as part of their daily workflow.
Why enterprise could be the bigger prize
Consumer AI gets the attention because millions of people can download an app overnight.
Enterprise AI works differently.
Businesses care about whether a product can save employees time, reduce repetitive work and integrate with the systems they already use.
That makes integrations particularly important.
An agent that can work across Slack, calendars, project-management tools and other business applications has a much larger potential role than an AI that simply answers questions.
Imagine an employee starting the day with dozens of notifications.
Instead of manually sorting through everything, an agent could potentially identify what needs attention, investigate routine issues, prepare responses and escalate the decisions that actually require a human.
The value is not just intelligence. It is execution.
That is where the agentic AI opportunity gets interesting for businesses.
Meta is moving in the same direction
The competition is already expanding beyond consumer apps.
Meta has announced that Muse is being expanded to small businesses, with integrations including Shopify, Dropbox and Slack, alongside a broader list of business and productivity tools.
Muse can also connect with Meta’s own business ecosystem, including Instagram professional accounts, Facebook pages and advertising accounts.
That means Meta is no longer keeping its agent strategy purely in the consumer space.
Both companies are moving toward businesses. They are simply arriving from different starting points.
Meta has an enormous consumer and social ecosystem.
OpenAI has ChatGPT and a growing developer and enterprise footprint.
The battle is increasingly about who can build the most useful layer between people and the software they already use.
But there is a major catch: autonomy comes with risk
The more capable AI agents become, the more important their mistakes become.
A chatbot giving you the wrong answer is one problem.
An autonomous agent taking the wrong action is another.
The source material points to a growing number of incidents involving AI agents behaving in unexpected ways, including cases involving organizations and government websites.
That is why the AI industry is facing a difficult balancing act.
Companies want agents that can act independently.
But the more independence they give them, the more carefully those actions have to be controlled.
OpenAI has also faced safety concerns around its latest models, while Anthropic CEO Dario Amodei has called for frontier AI development to slow down.
That tension is unlikely to disappear.
The commercial opportunity for agents depends partly on making them autonomous enough to be useful, but controlled enough that businesses are comfortable giving them access to sensitive systems and information.
The investor question is bigger than Dots
For investors, Dots itself is only one part of the story.
The bigger question is whether agentic AI becomes the next major monetization layer for the AI industry.
The generative AI boom created enormous demand for computing, models and infrastructure.
The next phase could shift more of that value toward applications that actually perform work.
If businesses begin paying for AI agents that handle meaningful portions of their workflows, the economics of AI could change again.
That could create opportunities across:
- AI model providers
- Cloud and computing companies
- Enterprise software
- Cybersecurity
- Data infrastructure
- AI application developers
But there is also a risk.
If customers become more price-conscious and increasingly use multiple AI models rather than relying on a single provider, AI companies may face pressure to prove that their products deliver measurable value.
Being the smartest model may not be enough anymore.
The winning product could be the one that saves a company the most time, handles the most work and makes the fewest costly mistakes.
OpenAI is also playing a much bigger financial game
Dots arrives as OpenAI continues to scale rapidly.
The source material reports that OpenAI’s annualized revenue was approaching $70 billion, while the company is reportedly preparing for another major funding round at a valuation that could reach around $1.4 trillion.
At the same time, OpenAI has been linked to plans for a potential public offering next year.
That makes product execution increasingly important.
At enormous private-market valuations, investors will eventually want to see more than user growth and excitement around new models.
They will want evidence that AI can become a durable, profitable business.
Enterprise agents could be one of the clearest paths toward that goal.
Businesses are already accustomed to paying for software that saves employees time or automates work.
If AI agents can become part of that spending, the market could be significantly larger than today’s chatbot economy.
The next battleground: teams of AI agents
Perhaps the most interesting part of OpenAI’s Dots announcement is where the company says it wants to go next.
Users start with a single Dot that they can name and customize.
But OpenAI envisions a future where users could have entire teams of Dots working together.
That takes the concept much further.
One agent could monitor incoming information.
Another could handle research.
Another could work through a project-management system.
Another could prepare reports.
And a human could remain in charge of approvals and decisions.
If that vision works, the workplace relationship with software could change dramatically.
Instead of employees constantly moving between applications, AI agents could become the layer connecting those applications together.
The AI race is entering a new phase
The first phase of the AI boom was about getting machines to generate.
The next phase is increasingly about getting machines to act.
OpenAI’s Dots, Meta’s Muse and the broader push toward agentic AI all point in the same direction.
The technology is moving from:
“Ask me anything.”
to:
“Tell me what needs to get done.”
That is a much bigger promise.
It is also a much harder one to deliver safely.
For investors, the opportunity is enormous, but so are the questions around reliability, pricing, regulation and trust.
The companies that win this race may not necessarily be the ones with the most impressive demos.
They may be the ones that convince businesses to hand over real work.
And that is why Dots is more than another AI product launch.
It is another sign that the AI competition is shifting from intelligence to execution.