MCP is not just hype. Here’s why I care and you should too:
I’ve been using Zapier for years for my own automation at LearnDataEngineering to automate processes like student enrollments and messaging. But I always saw it more for business processes than data engineering.
Well, they now have MCP servers and a SDK which makes Zapier super interesting for us engineers. Because we can use the thousands of platform connectors that they already have.
This is important because:
→ MCP lets a model like Claude talk directly to tools like Databricks, Google Drive, Gmail. Think about chat interfaces
→ The SDK gives us direct code access to these thousands of integrations and actions
→ Zapier handles the authentication to all the platforms. So, less annoying work for us
→ You can combine information across all these platforms
We already started to develop a real-world Databricks data quality use case that solves a problem many people have. Chat interface, dashboard, email alerter and more.
→ Understand the difference between MCP and the SDK and which one fits how you build
→ Read the SDK guide to understand how to give AI agents access to your app stack
→Watch a full workshop replay on building a real agentic workflow with Zapier from scratch
→ and more.
