Interview

Monogram launches voice-first AI app with visual UI output, raises $40M from DST and Lux Capital

Jun 30, 2026 with Eren Bali

Key Points

  • Monogram, a voice-first AI app that generates visual interfaces instead of text responses, launches publicly with a $40M seed round co-led by DST Global and Lux Capital.
  • The app delivers AI-generated UI layouts within 1.5 seconds by running 20 to 30 parallel processes, using OpenAI's model with Monogram's own fine-tuned layers on top.
  • Founder Eren Bali targets everyday search and discovery use cases—movie recommendations, restaurant lookup—positioning Monogram as the consumer-accessible entry point to AI after years of enterprise and developer focus.

Eren Bali has quietly built and launched Monogram, a voice-first AI consumer app, announcing a $40M seed round co-led by DST Global and Lux Capital on the same day.

What Monogram does

The core idea is straightforward: instead of returning text, Monogram generates a complete visual interface in response to a voice query, typically within 1.5 seconds. Bali frames this as the natural evolution of how humans process information — voice input is roughly four to five times faster than typing, but listening back to audio is slow. Visual output lets users absorb and interact with information faster than either. Tapping a card for more detail is quicker than asking a follow-up question out loud.

The app is currently iOS-only, with Android, desktop, and web clients described as in progress. Because the entire interface is generated by AI, Bali argues the app is inherently cross-platform once clients are built.

It's a general purpose AI application for everyday use. With other AI applications, you usually have a chat based interface. With Monogram, you ask something and then you get a complete visual interface back. We generate a visual interface response in roughly one and a half seconds. The round was led by DST and Lux. It was a $40,000,000 seed round.

How it works under the hood

Bali describes an architecture where one primary language model handles the response, while roughly 20 to 30 parallel processes run simultaneously to handle information architecture, visuals, and interactivity. The underlying model is OpenAI's, with some of Monogram's own fine-tuned models layered on top.

He's watching diffusion models closely as a potential replacement for the main model, arguing the architecture is conceptually right for what Monogram is doing, but says diffusion-based models aren't yet accurate enough for production use. Pixel-rendered interfaces — think real-time diffusion generating what looks like a website — are something he rules out as the primary form factor in the near term.

Beachhead use cases

Monogram isn't targeting enterprise workflows or coding agents. The early traction is in everyday search and discovery queries: movie recommendations, restaurant lookup, simple open-ended questions. Bali sees this as the natural entry point, the same instinct he brought to Udemy (learning Photoshop in a week, not online degrees) and Carbon Health (accessible care, not premium concierge medicine).

The throughline across all three companies is making something complicated accessible to more people. Monogram's version of that thesis is that AI has mostly served developers and enterprise users, and a simpler, faster consumer interface unlocks a much larger population.

The $40M seed at launch, led by DST and Lux, is a significant bet on that thesis finding distribution before the major platforms close the gap.

Every deal, every interview. 5 minutes.

TBPN Digest delivers summaries of the latest fundraises, interviews and tech news from TBPN, every weekday.