Eren Bali's Monogram raises $40M for a visual AI app
The Udemy and Carbon Health co-founder is betting that AI's next bottleneck is the interface, not another foundation model.
By Ryan Merket · Published
Why it matters
Monogram is a bet that consumer AI's next company can be built at the interface layer, where user intent, context and trust may matter as much as model quality.
Eren Bali (@erenbali), Edouard Tabet (@edouardtabet) and Murat Akbal (@akbal_m) launched Monogram, an iPhone app that turns AI responses into interactive visual interfaces, and announced a $40 million seed round led by DST Global and Lux Capital.
Bali framed the launch in a July 7th post on X as Monogram's exit from stealth. The stronger signal is in the June 30th company post: Monogram is entering the crowded consumer AI market without claiming to own a frontier model. It is trying to own the surface area where ordinary users ask AI to do things.
That is a specific founder bet. Bali built Udemy around access to instruction, then Carbon Health around software wrapped around a clinic network. Monogram pushes the same product instinct into AI: the model layer may be powerful, but the user still has to fight a text box. Monogram's answer is a general-purpose AI app that generates a user interface in seconds, then lets the user interact with that result visually.
Monogram says the app is designed for everyday tasks such as finding something to watch, making a recipe, planning a trip, finding a restaurant or comparing EVs. The App Store listing describes Monogram as an app that answers questions with a visual and interactive response instead of a chatbot thread. The app is free, iPhone-only, listed in the Lifestyle category, and requires iOS 26.0 or later.
A seed round for the application layer
The financing was led by DST Global and Lux Capital, with participation from Conviction, SOMA Capital, Gradient Ventures, e2vc, Maxitech and a group of angel investors that Monogram says includes Arthur Mensch, Logan Green, Karim Atiyeh, Garry Tan, Pete Flint, Mario Schlosser, Othman Laraki and Lenny Rachitsky.
Monogram did not disclose a valuation. It also did not disclose usage, revenue or retention. That leaves the public story closer to a product thesis than a traction story. The $40 million seed gives Bali and his co-founders enough capital to test whether consumer AI distribution can still be won at the interface layer, where most challengers face a hard problem: ChatGPT, Gemini, Claude and Perplexity already sit on user habits, model access and brand recognition.
Monogram's distinction is that it treats the chat window as technical debt. Its October 1st manifesto compares AI's current interface to earlier pre-mainstream computing moments: the command line before graphical interfaces, the internet before browsers, and phones before multi-touch. Strip away the historical sweep and the product argument is direct: AI needs a visual UI, privacy controls and collaboration features if it is going to move past prompt-and-response sessions.
The privacy test comes early
Monogram's own terms make clear how much user context the product may need. The June 14th terms of service say the app is in beta, uses Google Sign-In as its sole authentication method, and requests basic Google profile information plus read and write access to Google Calendar so Monogram can schedule and edit events and incorporate calendar events into AI outputs.
That access fits the product direction. An assistant that plans a birthday dinner, suggests restaurants or builds a trip itinerary becomes more useful when it knows time, location and preferences. It also pushes Monogram into the trust problem that consumer AI apps keep running into: a more personal interface requires more personal data.
The App Store privacy label says Monogram may collect location, contact info, identifiers and usage data linked to the user, along with diagnostics not linked to the user. Monogram's terms say users retain ownership of content they submit and AI outputs they receive, while also granting Monogram licenses to use user content and AI output to operate and improve the app. The terms carve out Google API data from model training and say use of individually identifiable data for model training is subject to prior consent.
That language matters because Monogram's pitch depends on trust as much as taste. In its manifesto, Monogram says AI needs to remember what matters while redacting, encrypting and keeping the user in control. The legal terms draw the practical boundary around that claim: the product is still a beta app using third-party AI services, and Monogram says AI outputs may be inaccurate, incomplete or flawed.
Bali's third act is an interface company
Bali's founder history gives Monogram instant attention and a tougher benchmark. He co-founded Udemy, then Carbon Health, and now presents Monogram as a new type of software rather than a vertical workflow tool. His LinkedIn profile describes his current work as "inventing a new type of software at Monogram," a phrase that matches the ambition of the launch more than the current evidence.
The current evidence is an iPhone app, a large seed round and a clear product opinion. Monogram is betting that the winning consumer AI interface will feel less like asking a search box for help and more like receiving a custom mini-app for the task at hand. If that works, the value accrues to the layer that owns user intent and interaction. If it fails, Monogram becomes another wrapper fighting default behavior set by the model companies.
The funding buys Bali's team time to prove which side of that line Monogram sits on.