AI is changing the data landscape fast. Amongst the obvious capability changes (like speed and ease of coding) is a subtle reframing of the methods and types of data that we can now access and turn into insights. Data is no longer just rows and numbers it's support calls, documents, strategy decks, and the context scattered across all the tools your business runs on. In particular there are three big shifts:
- MCP servers are becoming the universal connector. They give you a single, standard way to plug your business together, and open the door to data from unconventional sources and the ability to build AI-powered workflows.
- Access to business context becomes a huge lever of value. The more AI knows about your business, the more relevant and insightful its answers become.
MCP powers
We want Count to be the best place for you to explore and understand your data and from today you can now connect MCP servers as a data source in your canvas. This gives you new ways to access data and context from across your business and data stack.
Structurally MCP servers behave very similarly to existing database connections or catalogs: Once you’ve added an MCP connection you can then assign them to specific projects in Count, giving every user in that project the ability to pull in information from that connection into the canvas using Count’s agent.
We've built custom integrations for 24 of the most popular MCP servers but you can connect any other using the generic MCP connection.
To learn more about how MCP connections work and how to get started check out 👉 our docs 👈
A few of our favourite use cases
This opens up a huge range of possibilities, and we'd love to see what you build. Here are a few to get you started:
- Analyse qualitative data alongside quantitative data. Build a customer 360 view that combines usage data with the themes and trends pulled from support calls or messages.
- Bring in business context from your company documents. Analyse data in the context of your latest feature releases, or give AI access to strategic priorities and agent skills living in other tools.
- Combine trusted data from across your stack. Connect a third-party semantic layer and blend it with data already in Count.
What's next
This feature is launching in beta, and we'll keep expanding the data and capabilities you can use with MCP servers in Count. .
A few weeks ago we launched our MCP client, so you could take Count's agent and analysis into your other tools and workflows. Now, with MCP connectors, Count works in both directions - fully interoperable with other tools in your business.
Your data has never really lived in one place and now it doesn't have to. By treating MCP servers as first-class data sources, Count becomes a tool where you can not only explore more of your data but do so in new ways as well. We can't wait to see what you find!
