How Miro is trying to make AI more collaborative

For all of AI’s copilot branding, people tend to fly solo in many of their interactions with the technology: one human prompting various chat terminals.

Collaboration platform Miro wants to toggle the switch to “multiplayer.” The workflow visualization platform has rolled out a series of new AI tools—called Flows and Sidekicks—that are meant to let a whole team interact with the technology at once. The company debuted new AI Workflows this week.

Miro’s not the only company thinking along these lines: A few months ago, a Microsoft exec hinted at tools that make AI more of a “team sport,” per GeekWire; OpenAI has been experimenting with “multiplayer ChatGPT.” Indeed, “multiplayer AI” seems to be hovering on the verge of buzzword-dom.

We spoke with Miro Chief Revenue Officer Norman Gennaro about how the company is thinking about this mode of AI interaction, whether businesses are actually ready for it, and how burgeoning communications protocols play a role.

This conversation has been lightly edited for length and clarity.

Can you explain the concept of multiplayer AI: What sets Miro—and the way it’s using AI—apart from some of these other enterprise platforms that are more focused on individual use?

Miro is a collaboration platform. And I think what we see in most AI tools today is they are what we would call “single-player”...You use something like OpenAI, you put a prompt in, you get an answer out. You may cut and paste it into other services and go there. Given that we believe collaboration is so key, Miro allows you to collaborate on a canvas and then use the canvas as a prompt into the AI, getting an answer back that takes into account what everyone who’s been collaborating is thinking about. So we firmly believe that collaboration is important, and when you bring it into a multiplayer environment, enabling [you] to take those cues and put that into the prompt, you get a better answer at the end of the day.

It seems like a tricky balance to have an AI that’s not dominating the conversation, but rather collaborating without becoming extraneous. How did you make sure that the AI was within this collaboration rather than outside of it?

What we really try to do is not pass it off solely to AI, but have the human in the mix. So you’ll see us doing it in two ways. You’ll see our Flows, which enable you to take a process from one step to the next to the next, insert the human in the middle. But you also see us with Sidekicks, and the Sidekick is able to watch what you’re doing and give you advice as you go. So these two—I wouldn’t say competing—but two different ways of looking at the problem at hand allow you to get an AI that is collaborating with you, almost like a peer set.

In terms of the companies that are using Miro, what does a company need to do to be ready for multiplayer AI? How do you break down the silos between teams, or make sure that it’s in a place where this makes sense?

So if we think about our most advanced customers who are using Miro, what they’ve done is kind of looked at the flow that they’re trying to run, and they’ve broken it into pieces so that they can have the collaboration throughout the entire set. And the things that they have to prepare for, if they haven’t done that before, one is just getting people comfortable with AI in general—which is a thing that every corporation is working on right now, how to think through the AI transformation. But we’re fortunate in that Miro is a collaborative visual environment, so you’re looking at it and you’re leveraging AI as you go. So it’s very easy for people to kind of understand how AI can help them in a Miro environment, as opposed to a coding environment, or any of the other ways that you may do it in a single-player mode.

As you said, a lot of the AI out there is very much “single-player.” Do you have to break people out of thinking of AI that way? And has single-player almost given AI a bad name in some ways?

That’s an interesting way to put it. I think people know that it works in that single-player mode, but they want it to work in the multiplayer mode. So if you look at all the different tool sets that you have, regardless of the line of business that you’re in, you’ll see tools adding in AI all over the place. And then the question is, most customers are asking, “How do I stitch these things together? How do I start bringing it together for my entire flow?” And so it’s actually less of us having to explain to people, “Hey, this is how we think it should work,” and more of them saying, “I wish it worked this way.” And our answer is, “Let us show you how you can do that inside of Miro.”

With frameworks like Anthropic’s Model Context Protocol (MCP), are we starting to see more of these barriers between enterprise tools breaking down?

Yeah, so MCP is a core integration piece. And you’ll see that inside of Miro there is an MCP server in beta today. I’m not sure when we’re saying we’re actually going [general availability] but that’s very soon…Think of Miro as the canvas in which you’re leveraging all of the tools—we bring in OpenAI, Anthropic, any of the others—but you’ll see some of our largest enterprises incorporating MCP to make additional connections into Copilot, into Lovable…Cursor. But also import things. So if they have an MCP server, they can connect to ours and start bringing this onto the canvas, so that your entire workflow and your process that you’re trying to work through is on the canvas itself. So it’s both input-output, and think of it as a circular way that you can continue to do in advance, but also bring it back into Miro.

MCP has caught on really quickly, but it’s still early days. How would you like to see it evolve? What do you still need from that integration perspective?

MCP is making it incredibly easy, and almost everyone at this point is, if they don’t have an MCP server, they will shortly, and they’re looking at that. So it’s a key integration. But I think most of the software out there, we really want to think about it as a two-way interaction. So how do I get something from MCP into something like the canvas, and then how do I return it back so I can have the full loop? And that’s where we’re spending a lot of time.

Do you think we’ll see more multiplayer AI? Do you think it’s going to be a big step change in the coming year? What do you see as kind of the future of this trend?

Once people understand what we’re doing—there’s not very many people that are doing this right now…I think it’s unique in the market right now. But I think what we’re hearing from executives, from leadership, from other people inside of these companies [is that] this is the way to bring the entire enterprise to the AI transformation, understand it, leverage it, and make use of it. I’m pretty confident that this will be a long-term approach of how to take AI and humans, bring them together, put them in a space where they can understand what’s working, how it’s working, and collaborate.