2.6 Connecting Data Sources

🎥 Video lesson — video not yet migrated (source Skool video id: a337c00d41ba4d60b1719390065da2ef).

Learning Objective

Understand the four ways to connect external data to your Business Brain and successfully connect at least two data sources that Claude can access.

Why You Need Data Connections

Your Business Brain has context and process maps, but it cannot actually do much without access to real data. Email, spreadsheets, CRM, databases - this is where your business actually lives.

Connecting data sources is what turns Claude from an advisor who gives good answers into an operator who takes real action on real information.

Option 1: API Skills (Recommended)

An API skill calls an external service only when you need it. It is clean and efficient because it only loads data when triggered, not every conversation.

For example, you could build a skill that pulls leads from your CRM only when you ask for them. The rest of the time, it takes up zero space. This is the best approach for any tool that has an API, which is most modern tools.

Option 2: MCPs - Model Context Protocol (Use Sparingly)

MCPs give Claude live, always-on access to external tools like Gmail or Google Sheets. They are powerful but they eat your context window because Claude loads the connection info in every conversation.

Use MCPs for tools you access constantly throughout the day. Avoid them for tools you only check occasionally. If you find yourself relying heavily on an MCP, that is a sign you should convert it to an API skill instead.

Option 3: Manual File Drops

This is the simplest approach. Export data from a tool, save it as a file in a folder Claude can read. No setup, no configuration.

This works well for one-time imports or data that does not change often. For example, export your client list as a CSV and drop it in your context folder. Claude can read it whenever needed.

Option 4: Scheduled Syncs

Scheduled syncs are automated scripts that pull fresh data on a schedule - daily, hourly, whatever you need. They run in the background and keep your data current without you doing anything.

This is the best approach for dashboards, reports, and metrics that need to stay up to date. It is more advanced, and you will learn how to build these in Part 3 (AUTOMATE).

How to Choose the Right Option

The rule is simple.

If you access it multiple times per day, use an MCP. If you access it for specific tasks, build an API skill. If the data rarely changes or it is a one-time import, do a manual file drop. If you need it always fresh in the background, set up a scheduled sync.

Start with MCPs because they are the easiest to set up. As you get more comfortable, convert your most-used MCPs to API skills for better performance.

Walkthrough: Connecting Common Sources

Gmail is one of the most useful connections. Add the Gmail MCP to your Claude Code settings and you can ask Claude things like "Show me my last 5 unread emails" or "Find all emails from this client." This is essential for your daily brief and any email-based workflows.

Google Sheets is another high-value connection. Set up the Google Sheets MCP and Claude can read your spreadsheets, pull specific data, and work with your client trackers, content calendars, or lead lists directly.

CRM or database connections follow the same pattern. Whether you use HubSpot, Notion, Airtable, or a database like Postgres, you give Claude access and then it can read and act on your data. Check the course resources for setup guides specific to your tools.

Action Items

  • Pick 2 data sources your business depends on daily (email, spreadsheets, CRM, etc.)
  • Connect them using MCPs as your starting point (setup guides are in course resources)
  • Test each connection by asking Claude to pull real data from each source
  • If a connection does not work, check the course troubleshooting guide or ask in the community
  • Write down which sources you want to convert to API skills later (you build those in Part 3)
← Previous
Next →