Author
Henrik Wohlatz Llebot
CRM Developer & Consultant
Table of contents
Share this article now!
Author
CRM Developer & Consultant
Many teams are already using individual AI agents, often without an overview of what each one is doing or whether they're aware of each other. This is precisely where HubSpot's Agent Hub comes in. It brings all agents for marketing, sales, and service into one place, with a shared context and a console for status and performance. This article shows which agents are available, which one is right for which team, and how to set them up step by step.
The most important information at a glance
Many teams already use individual AI agents, often without an overview of what each one does or whether they are aware of each other. This is precisely where HubSpot's Agent Hub comes in. It brings all agents for marketing, sales, and service into one place , with a shared context and a console for status and performance. This article shows which agents are available, which one is suitable for which team, and how to set them up.
The Agent Hub has been renamed and is still new. It launched as a public beta on July 23, 2026, and is available to customers with a Professional or Enterprise plan. Instead of five separate AI tools, each learning your business from scratch, agents here share the same context drawn from your CRM data, brand language, and target audiences.
The Agent Hub is HubSpot's central location for AI agents that autonomously handle tasks in marketing, sales, and service. These agents were previously known as Breeze Agents. A console allows you to see which agents are active, review their performance, activate additional agents, and build your own using the Agent Builder.
While individual agents have existed for some time, they previously operated independently. This led to the situation HubSpot itself describes as problematic: a prospecting agent might contact a lead while, simultaneously, a customer agent handles an open complaint from the same customer, without either being aware of the other's activity. With the Agent Hub, all agents now converge in a single interface and access a shared context. The former umbrella term "Breeze Agents" is therefore no longer used. The name remains only for the Breeze Assistant, the interactive AI helper for daily tasks, which is distinct from the autonomous agents.
The Agent Hub currently comprises five functions: the Customer Agent for service, the Prospecting Agent for sales, Deal Progression for managing the pipeline, the Data Agent for clean CRM data, and AEO for visibility in AI searches. In addition, there's the Agent Builder for creating your own agents tailored to your business.
The Customer Agent answers frequently asked customer questions independently and around the clock via channels such as chat, email, and messenger. They draw on your knowledge base and the contact and contract data in your CRM to resolve straightforward cases independently. If a request becomes too complex, they escalate it, along with its history, to a team member. This agent is ideal for service teams with many recurring questions who want to streamline routine tasks and free up more time for complex cases.
The Prospecting Agent takes over the time-consuming preparatory work in sales. They monitor the CRM for buying signals such as funding rounds, product launches, or new job openings, identify suitable contacts, and prepare personalized outreach. Since the system was restructured in spring 2026, they cover the entire prospecting process, from research and building contact lists to drafting messages and booking appointments. They are ideal for sales teams that want more qualified conversations without spending hours on research.
Deal Progression helps sales teams move deals through the pipeline. The feature provides AI-powered recommendations on which deal needs attention next and what steps are advisable to ensure opportunities aren't missed. For teams with many deals running concurrently, this is a way to maintain an overview and focus on the deals that matter most right now.
The Data Agent takes care of the foundation upon which everything else is built: the data. It answers questions about your CRM data and helps keep records up-to-date and complete. Because every agent is only as good as the data it accesses, the Data Agent is especially valuable for teams whose CRM has become unwieldy over time.
The fifth feature is AEO, or Answer Engine Optimization. It tracks how often your brand appears when prospects ask questions using AI tools like ChatGPT and Gemini, and provides recommendations on what content you should create to be mentioned there. For marketing teams that want to be visible early in the research process, AEO bridges the gap between traditional search and the new answer engines.
The choice depends on the biggest bottleneck. A service team with many similar tickets is best off starting with the Customer Agent. A sales team that has too little time to engage new leads benefits from the Prospecting Agent, while teams with full pipelines are more likely to benefit from Deal Progression. If data quality is the core issue, the Data Agent is worth considering because it improves the foundation for all others. And marketing teams that want to appear in AI responses should start with AEO. In practice, you can start small and expand later because all agents share the same context and converge in the Agent Hub.
Setup is done centrally. In your HubSpot account, you'll find the Agent Hub in the navigation pane under Agents. There, you can see which agents are already active and activate a previously inactive one via Activate or adjust an active one via Manage. The shared context is crucial: In a dedicated area, you define your company message, tone, brand, and target audiences, along with knowledge repositories containing in-depth information such as product documents or internal guidelines. All agents draw their responses from this context, meaning you only need to maintain it once, not separately for each agent.
If the pre-built agents aren't sufficient, you can create your own with the Agent Builder. The builder is a no-coding interface where you describe in plain language what the agent should do and then define what it can access and which actions it is allowed to perform. Your custom agents work directly with the live data in your CRM, from deal history and contact data to buying signals, without you having to painstakingly map fields. This allows you to map workflows that are specific to your business.
A medium-sized company deliberately starts with a single agent. Because the service is struggling with the volume of standard questions, the team first activates the Customer Agent and integrates the knowledge base and brand voice into a shared context. After a few weeks, a large portion of the routine questions are automated, and the team moves on to the next bottleneck. They activate the Prospecting Agent for sales, which already uses the same context and therefore starts without any new setup. Later, the company builds a dedicated agent in the Agent Builder for a recurring, specialized task. This way, the deployment grows gradually, rather than trying to change everything at once.
Keep a few things in mind. The Agent Hub is still in beta, so features are evolving, and it requires a Professional or Enterprise plan. Pre-built agents and custom agents from the Agent Builder operate on a credit-based model, while with results-based agents, you only pay when work is actually completed, such as a resolved conversation or a recommended lead. Regularly review agent performance, especially at the beginning, and start with a clear use case instead of activating multiple agents simultaneously. This will help you maintain quality and oversight.
1. What is HubSpot's Agent Hub?
The Agent Hub is HubSpot's central hub for AI agents that autonomously handle tasks in marketing, sales, and service. They were previously known as Breeze Agents. In a console, you can view active agents, review their performance, activate more, and build your own.
2. What types of agents are available in the Agent Hub?
Currently, there are five: the Customer Agent for service, the Prospecting Agent for sales, Deal Progression for the pipeline, the Data Agent for clean CRM data, and AEO for visibility in AI searches. In addition, there's the Agent Builder for creating your own agents.
3. Are the agents still called Breeze Agents?
No. The agents now operate under the name Agent Hub. The former umbrella term Breeze Agents has been discontinued. Only the Breeze Assistant, the interactive AI helper for everyday work, retains the name Breeze.
4. Who can access the Agent Hub?
Customers with a Professional or Enterprise plan. The Agent Hub launched as a public beta on July 23, 2026, and is included in these editions. Agents' actual work is based on a credit-based model.
5. How do I set up an agent?
In your HubSpot account, you'll find the Agent Hub under Agents. There you can activate or manage agents. The shared context of brand tone, target audiences, and knowledge repositories is crucial, as all agents draw their answers from this.
6. Can I build my own agents?
Yes. With the Agent Builder, you can create your own agents without programming by describing what they should do in natural language. You work directly on the live data in the CRM and can tailor them to your business processes.
The Breeze Assistant is the practical AI helper for everyday work: It creates content in your brand voice, summarizes deals and tickets, answers questions about your CRM data, and performs small actions on command. Because it works directly within HubSpot and with real data, its answers are more relevant to your business than those of a generic chatbot. Maintaining your brand voice and data, asking specific questions, and reviewing the results will save you significant time. This transforms many small tasks into a smooth workday, leaving more time for actual customer interaction.
About the author
Henrik is a Customer Success and Operations professional with experience spanning technical B2B support, customer service leadership, and process development. He is passionate about building efficient systems, optimizing customer experiences, and creating structures that enable sustainable growth.
Support-Teams verbringen viel Zeit mit denselben wiederkehrenden Fragen, während die kniffligen Fälle warten. Kann KI das übernehmen? Für die häufigen Anfragen ja, rund um die Uhr und auf Basis...
Sie haben starke Produkt- und Vergleichsseiten, trotzdem empfiehlt die KI die Konkurrenz. Der Grund ist selten das Produkt, sondern die Form: ChatGPT, Gemini und Perplexity können die Seite nicht...
Your competitors are already appearing in AI-generated answers, but very few marketing teams know which questions they're answering and why. Those who fail to measure this are making incomplete...