HubSpot is at the heart of my customers' sales and marketing processes. But with the advent of generative AI, the rules of the game are expanding: Why search for information manually when an AI assistant can pull it directly from HubSpot? This is exactly where connectors and the Model Context Protocol (MCP) come into play. Two different ways to deeply integrate AI into your HubSpot usage. But which one suits your setup better?
Connectors are ready-made interfaces between ChatGPT and tools such as HubSpot, Google Drive or OneDrive. They allow ChatGPT to retrieve information from these applications in real time. They only access data for which you have regular authorization. In concrete terms, this means that you can ask ChatGPT "What is the current status of lead X in HubSpot?" and receive an answer with a source.
Important: Connectors are currently restricted to read-only access. ChatGPT can therefore retrieve and analyze information, but cannot write or change data in HubSpot.
MCP is an open standard developed by Anthropic to connect AI models to almost any external data source. Think of it like a universal adapter: Instead of building a special interface for each tool, your AI simply speaks "MCP" - and a corresponding MCP server takes care of the connection to the data source.
With MCP, you can not only read, but also execute actions: Create tickets, update contacts, trigger workflows. HubSpot can also be connected via MCP, especially when complex or individual processes are involved.
| Criterion | Connector | MCP |
|---|---|---|
| Simplicity | Plug & play in ChatGPT | One-time IT setup required |
| Data access | Read only | Read & write |
| Customizability | Limited to OpenAI offerings | Completely flexible, also for own systems |
| Security | Trust in OpenAI required | Full control, self-hosting possible |
| Area of application | Standard SaaS such as HubSpot, Drive | Also for special or legacy systems |
Breeze is HubSpot's own AI assistant. It uses HubSpot data exclusively and remains within the cosmos. This makes it strong when it comes to in-depth CRM understanding. But: Breeze cannot (yet) access external data sources and is functionally limited compared to ChatGPT plus Connector or MCP. So if you want to link HubSpot with other tools, there's no getting around Connector or MCP.
Connectors offer you a quick, uncomplicated introduction to the world of AI-supported HubSpot use. MCP, on the other hand, is your tool for long-term independence and in-depth process integration. Both technologies have their place - and when combined, they unfold their full potential.
Which HubSpot data would be particularly helpful for your AI? Are there any internal systems you would like to connect? Think about it and start with a pilot project. Because you can only learn if you start.