More than two million people have leveraged the Agent.ai platform developed by Dharmesh Shah, Founder & CTO at HubSpot.
Among them, 26,000 users built their own agents, and 1,800 of those agents are available for the community to use for free.
Shah shared the news at the company’s INBOUND 2025 event last week.
The Agent.ai platform is a network of autonomous AI agents for service, sales, marketing, and operations teams. Anyone can use the platform, even if they are not currently a HubSpot customer.
Users can discover, test, and deploy preconfigured agents within their CX ecosystem, and a low-code agent-building tool allows them to connect their own knowledge sources to the AI agent.
More and more users are doing so. Indeed, just 47,000 people had used the platform in September last year. By January, Agent.ai had 258,000 users, surpassing the 100,000 target that Shah set for the project.
Now, with 20x the users Shah had initially hoped for, the HubSpot man has great hopes for his company’s AI agent story. He said:
Last year at INBOUND 24, I predicted that this year would be the year of AI agents. I was wrong. This is not the year of AI agents, this is the decade of AI agents. We are just getting started on this massive transformational wave in AI.
Yet, despite the promise of AI agents and Shah’s initial aim for the Agent.ai platform to become part of HubSpot, it still runs independently.
Shah’s previous project, ChatSpot, did become part of the HubSpot ecosystem. It’s now “Breeze Copilot”.
Nonetheless, HubSpot has another group of AI agents: Breeze Agents. These are pre-configured, but customizable, and designed so that customers can drop the agents into their platforms. That’s ideal for many of Hubspot’s midmarket customers, which typically don’t have the skillset or time to configure AI agents from scratch.
The company released 18 new Breeze Agents in beta at INBOUND 2025 across marketing, sales, and customer service. These perform various tasks – such as automating customer contacts, prospecting, and customer research – and join the 15 Breeze Agents introduced previously.
With Agent.ai, the company may also use the platform to identify popular AI agents and leverage that intelligence to develop Breeze Agents that are effective and in demand.
However, as companies become more confident with AI agents, Agent.ai may come further into the picture and allow HubSpot customers to unlock new use cases and efficiencies.
TEAM: Shah’s Model for AI Adoption
In his presentation at INBOUND 2025, Shah also outlined a strategy to take AI from individual use to team adoption. He called it the TEAM strategy, with TEAM standing for:
- Triage – List all the potential AI use cases, then prioritize by frequency, impact, and low cost of failure.
- Experiment – Try, iterate, and refine. Adjust prompts, add context, and see what works.
- Automate – Once something works, build workflows or agents to scale efficiency.
- Measure – Track results. AI shouldn’t be performative; it should drive performance.
Via this model, companies can move from “individual heroics” to team habits that benefit the whole organization.
With HubSpot, like many other enterprise software companies, envisioning a future of human-AI hybrid teams, this model will help businesses get started.
