The Figma Make & user research handshake

One of the more time-consuming parts of the UX design process is the user validation phase; we have created a design we think solves a user problem, but we need to validate our assumptions. We conduct user interviews and then spend hours re-watching recordings, tagging transcripts in HeyMarvin (a research repository tool), and then trying to translate those messy insights into a clean UI in Figma. But recently, the workflow changed. With the release of Figma Make and its support for MCP (Model Context Protocol) servers, that middle step has been compressed from hours into seconds.

Figma Make + Hey Marvin

For those unfamiliar, a “connector” in Figma Make is essentially access to an MCP server. It allows Figma’s AI to reach outside the design file and “talk” to your other apps. In my case, I connected HeyMarvin, where all my qualitative research lives. Why is this worth reading about? Instead of analyzing the research in HeyMarvin and mining quotes to derive UX insights from, I can ask directly in Figma Make, the place where I’m building a working prototype of my design. I can use the HeyMarvin connector to reference my research directly on the canvas.

Steps to connect

1. Click the + sign in the Make prompt box.
2. Click “Connectors”
3. Find the connector of interest
4. Click “Connect” and follow the authentication prompts

You can find Connectors under the + sign button.

There are many connectors you can choose from to enhance your Make prototypes.

All you have to do to prompt is type: “@Marvin” and then tell it what to reference with your question. Here are some examples I’ve used recently:

· @Marvin reference [meeting title] with internal stakeholders and build a dashboard that would show them all the data they mentioned wanting to see for our newly released tool.

· @Marvin I remember a user from one of my recent interviews from [project name] complained about confusing error messages. Remind me what they said?

· @Marvin a user from [research session title] suggested a revised workflow for the prototype we showed them. Update the current prototype to reflect their suggested workflow.

The result is context-aware design

Prompting AI to create designs without user context can result in experiences that are “halfway there”. They’re on the right track but missing small details that matter to end users. When the AI has a direct line to the user feedback via the MCP server, it’s much more powerful. What I mean is:

· It’s able to identify the right KPIs: Figma Make no longer guesses what data should display on a dashboard, for example; it can pull the specific data points the stakeholders mentioned during user interviews.

· It can better map the information architecture: Make can now reference the hierarchy of needs expressed by the user.

· It’s generally more accurate and closer to the desired result: While you’ll still need to polish the AI output, it saves a lot of time in the “getting close to great” phase.

It did take me a minute to figure out exactly how to prompt Make to retrieve data from Marvin (i.e. referencing specific project and file names), so be prepared to wrestle with it a bit.

Why this matters for UX designers We’ve spent years talking about “data-driven design,” but this is context-driven design (something that is particularly important for getting better results when using AI tools). By using MCP connectors, we are no longer designing in a vacuum. We are designing with the context of our user research right inside our Figma files. This makes the design process less linear because user insight stays present deeper into the evolution of a project. Research is no longer something that only informs an iteration and then fades into documentation. With that context embedded directly in the design environment, it can continue influencing decisions throughout various increments of work. The process becomes more continuous, more integrated, and more grounded in actual user needs.

Have you tried Figma Connectors? If so, what has your experience been like?