Partnering with Tambo

Every AI assistant today works the same way: you type, it types back. Whether you're asking for a sales report or scheduling a meeting, the response is locked inside a wall of text in a chat window.

For enterprise software companies under pressure to add AI features, chat windows bolted onto existing products lack lasting power. Users try the feature once, get a wall of text and move on. 

What's needed then becomes AI that can ship features that actually look and feel like existing software vs. endless chatbots. That's why we partnered with Tambo. 

Building AI that speaks UI

Tambo was founded by Michael Milstead and Michael Magan, two builders who’ve spent years inside large enterprise software companies like Microsoft, Indeed, Convoy and Taxbit. They met at a hackathon, initially trying to solve the narrow technical problem of whether or not AI could generate website interfaces, not just text. 

The more they worked on it, the more a new belief emerged. If AI was going to fundamentally change how people interact with software, devs and product teams would need an entirely new toolkit to build those interfaces. 

At its core, Tambo is an AI agent that responds with interactive UI instead of plain text. Ask for sales by region and you get an actual chart you can click into. Ask to reschedule a meeting and you get a calendar, not a paragraph. The agent builds the right interface for what users are doing, so they can actually get work done. 

For developers, Tambo is an open-source toolkit that drops directly into the applications they’ve already built. It works with existing component libraries and design systems, which means teams don’t have to rip anything out or force users into yet another standalone chat experience. 

Why we partnered 

Generative UI is rapidly becoming consensus. Over the past few months, Anthropic rendered UI components in Claude through MCP Apps. Google released the A2UI spec. Vercel shipped json-render. 

But there's still a big gap between the spec and the product. Most teams can get an AI prototype working in days. Getting that prototype into production — with streaming, error handling, authentication, state management, and real UX — still takes months.

Tambo closes that gap. It’s the production-ready layer that turns generative UI from a concept into something teams can actually deploy, using the technologies they already rely on. Some early results from developers integrating Tambo into their existing stacks: 

• Product engineers at Solink went from discovery to production in a single weekend.

• Zapier’s AI agent team rendered their components exactly as they appear in their internal design system on the first try.

• Rocket Money’s product engineers integrated their existing UI components with a single line of code, eliminating significant custom work.

We backed Tambo’s seed with capital, and also partnered closely at key inflection points including collaborating on technical architecture, early team-building, and launch execution to help the company compress timelines as they scaled (h/t Alec and Jose). 

We're proud to be early partners with Magan and Milstead as they build the interface layer for AI-native software. Check out Tambo here.