🔥 Feature Spotlight: Add MCP Tools to Make AI Agents

News
, , ,
1

Your Make AI Agent can now connect to external MCP servers :electric_plug:

If you are building AI Agents that need to act across more of your tools, this update opens up a lot of new room to build.

Make AI Agents can now access tools exposed via external MCP servers and use them within the same run as native Make modules.

:electric_plug: Introducing MCP tools for Make AI Agents: add external MCP servers directly in the Make AI Agent builder and give your agent access to the tools those MCP servers expose.

See it in action :backhand_index_pointing_down:

:new_button: What’s new with MCP Tools for Make AI Agents

  • :electric_plug: Add external MCP servers to AI Agents: connect one or more MCP servers as tool providers from the AI Agent builder.

  • :toolbox: Give agents a larger toolset: your agent can read the tool list exposed by each MCP server and decide when to use those tools during a run.

  • :high_voltage: Mix MCP tools with native Make modules: agents can call MCP tools and Make modules in the same reasoning loop, based on the instructions you give them.

  • :globe_with_meridians: Use a standard protocol: connect to any standard MCP server, so teams can bring existing MCP setups into Make and build broader agent workflows faster.

:flexed_biceps: What to automate with MCP Tools and Make AI Agents

  • Sales lead research: trigger an AI Agent when a demo request comes in, use MCP tools to research the company, enrich the contact, and write a structured summary back to the CRM before the rep starts their follow-up.

  • IT support triage: let an agent check internal docs through an MCP server, compare the ticket against recent code or product updates, draft a first response, and update the ticket in one run.

  • Weekly market briefing: have an agent scan selected web sources through MCP, compare findings with your internal roadmap or notes, and post a clean summary to the right channel before the Monday standup.

This is especially useful for AI agent builders, automation engineers, developers, and agency teams building agents that need to work across client tools, internal systems, and external data sources.

Connect the MCP server once, define the agent instructions clearly, and let the agent choose the right tool for the job.

image
image1170×723 55.3 KB

By adding MCP to Make AI Agents, you get the structure and reliability of workflow automation, along with the reach of the open MCP ecosystem. One agent, one flow.

Demo time

What will you connect to your Make AI Agent first?

Happy automating! :purple_circle:

Valery from Make

:books: Helpful resources

:backhand_index_pointing_right: To learn how to connect your agent to MCP tools, see Create your first AI agent

5 Likes
:purple_circle: Releases on Make: June 2026
2

Great addition, Valery. We have been building lead response automations in Make for clients for a while now and this is one of the updates we were genuinely excited to see land.

The first question it raised for us was whether to rebuild our existing lead response scenarios as AI Agents now that MCP tools are in. We worked through it carefully and landed on no, and the reasoning might save someone here some pain. Agents reason, pipelines execute. The deliberation an agent does (read the goal, weigh the tools, decide what to call) is exactly what you want for research and follow-up work, and exactly what you don’t want in the first 60 seconds after a form submission. Lead response needs the same steps in the same order, in seconds, every time. That stays a deterministic Make scenario.

Where MCP agents slot in cleanly is everything after the first reply goes out: deep company research beyond a basic enrichment lookup, drafting the follow-up sequence, checking a rep’s calendar through an MCP server before proposing times, watching a thread and nudging a rep who went quiet. None of that is time-critical, and all of it benefits from a model deciding which tool to reach for. Make is the right place to build both sides of this, which is exactly why we keep coming back to it.

The test we use now: if the task has a deadline measured in seconds, build a Make pipeline. If it benefits from reasoning across tools and a minute of thinking costs nothing, give it to a Make AI Agent. MCP didn’t move that line. It just made the agent side of it far more capable.

We wrote up the full lead response architecture built in Make, including a new section on exactly where MCP agents fit and where they don’t: How to Build an AI SDR That Replies, Qualifies, and Routes Every Inbound Lead in Under 60 Seconds - Rivetline Consulting - AI Search Optimization and CRM Speed to Lead Optimization

Curious what others are connecting first. We’re eyeing our WordPress MCP server for a content workflow. That feels like the natural agent-shaped job.

1 Like