Metabase integration with lemlist through MCP, with prompts to copy

Your data team built a dashboard in Metabase that flags the top free workspaces by usage, the trial accounts about to hit a limit, the customers matching an expansion signal. Someone glances at it in a weekly meeting. Nobody exports it, nobody loads it into an outreach tool, and the list goes stale before a rep ever sees it. That gap between a saved question and a live campaign is where most pipeline opportunities quietly die.

This article gives you a metabase lemlist integration built on MCP, with prompts you can copy into Claude, ChatGPT, or Cursor today. I’ll show you why connecting both tools through MCP matters, what you need to set it up, and two concrete workflows: turning a saved question into a campaign, and checking which campaigns actually produced paying customers. You’ll also get the guardrails to keep an agent from making a mess in either tool.

By the end, you’ll know how to turn a dashboard into pipeline without writing a single line of export logic.

Why connect lemlist and Metabase through MCP

Your best target lists already live inside Metabase as saved questions. When both MCP servers are connected in the same AI client, an agent can run a Metabase question and turn the result into a lemlist campaign in one conversation, no CSV export, no copy-paste between tabs.

An MCP server is a lightweight layer that exposes a tool’s actions to an AI agent through a standard protocol. The agent calls those actions the same way you would use the tool’s UI, but it does it from a single chat window. If you want a deeper explanation of how this works for outbound, read the guide on MCP for sales.

With a Metabase MCP server, the agent can run saved questions, browse dashboards, and query databases. With the lemlist MCP server, it can search leads, enrich contacts, build sequences, and launch campaigns. Put them together and the agent reads your BI data, then acts on it in lemlist without you switching tools.

What you need before you start

• •

  The lemlist MCP connected in Claude, ChatGPT, or Cursor. Follow the walkthrough on how to set up the lemlist MCP if you haven’t done this yet.

• •

  The Metabase MCP connected in the same client. Your Metabase admin turns on the MCP server under Admin > AI > MCP, and you point your client at the endpoint.

• •

  Permissions in both tools scoped to what the use cases actually need. The Metabase MCP respects the connecting user’s permissions, and the lemlist MCP follows your plan and workspace role.

Start with read-only prompts. Let the agent pull data from Metabase and read campaign stats in lemlist before you give it permission to create or update anything. Once you trust the output, move to write actions.

1. Turn a saved question into a campaign

Data teams build great lists inside Metabase. Free workspaces sorted by usage, trial accounts approaching a limit, customers who match an expansion signal. Those lists sit in a dashboard and never reach the rep who should act on them. If you are still defining which accounts fall inside your total addressable market, start there before pulling a list.

Copy this prompt into your AI client with both MCPs connected:

Here is what the agent does. It calls the Metabase MCP to run the saved question and return the results. It parses the rows for company names, domains, or contact details. It then calls the lemlist MCP to search the People database for matching contacts, enriches them with verified emails through waterfall enrichment, and imports them into a new campaign.

Before you run this on a full list, check the results the agent pulled from Metabase. Confirm the column mapping is correct, verify the enrichment matched the right people, and review the campaign draft. If you are unsure about batch size, the guide on how many leads per campaign gives a practical framework.

2. Check campaign results against revenue

Most outbound reports stop at opens, clicks, and replies. You know which campaign got the most interest, but you don’t know which one produced paying customers. That answer lives in Metabase, in a table your finance or product team already maintains.

Copy this prompt:

The agent reads your lemlist campaign stats and exports the lead list with company names and domains. It then runs a query or saved question in Metabase that returns paying accounts. It matches the two lists by company domain or name and tells you which campaigns converted, which ones generated pipeline that stalled, and which ones produced nothing.

Before running this across every campaign, confirm the agent is matching on the right field. Domain matching is more reliable than company name. Ask the agent to show you the matched pairs before it draws conclusions.

Guardrails when an agent works in lemlist and Metabase

Giving an agent access to two tools means twice the surface area for mistakes. These guardrails keep things tight.

• •

  Review every campaign the agent creates before you launch it. The agent can draft sequences and import leads, but a human should approve the final version.

• •

  Test on 10 leads first. Run the full workflow on a small batch, check the data, then expand.

• •

  Keep unsubscribes and exclusion lists active in lemlist. The agent respects these, but only if they are turned on.

• •

  Decide who on the team can connect the MCPs. Limit access to people who understand both tools and the data flowing between them.

• •

  Log what the agent changed. After each session, ask the agent to summarize the actions it took. If something looks wrong, you can trace it back.

Run it in lemlist

The lemlist MCP server works in Claude, ChatGPT, and Cursor. Once connected, the agent has access to the full outbound workflow from a single chat.

It can search the People database with 650M+ contacts across 65M+ companies. It can enrich leads with waterfall enrichment across 14+ providers to find verified emails and phone numbers. It can create campaigns and sequences across email, LinkedIn, calls, and WhatsApp. It can read the unified inbox and campaign stats to tell you what is working. And it can act on intent signals to prioritize leads showing buying behavior right now.

These are the same capabilities covered in the full list of lemlist MCP use cases. The Metabase connection adds a data layer on top: instead of starting from a blank search, you start from a question your BI team already answered.

Final takeaway

Your BI tool already knows which accounts deserve outreach. The lemlist MCP and a Metabase MCP, connected in the same AI client, let an agent turn that knowledge into a live campaign in minutes. The gap between insight and action shrinks to a single prompt.

Build your first signal-based sequence in lemlist.