ElevenAI: the supplier behind meeting-note apps now competes with them

Picture your best supplier opening its own shop one street over, selling the same goods. That is what happened at the ElevenLabs Summit in New York on November 11, just with software.

What ElevenLabs announced

• ElevenAI: a new app for employees that turns conversations, including in-person meetings, into notes and next steps. Dictation and meeting notes run on Scribe v3, ElevenLabs' latest speech-to-text model.
• Consent built in: according to ElevenLabs, the app only takes notes in a meeting when everyone agrees, then sends every attendee a summary and action items.
• Availability: free early access since November 11. The announcement does not mention pricing, languages or platforms.
• Around it: memory for returning customers and pre- and post-call routines for its customer service agents (ElevenAgents), plus a partner program with more than 550 partners in 44 countries.

It looks like a feature. It is a change of sides

Until now ElevenLabs was the supplier. Others built the app, ElevenLabs supplied the ears. ElevenLabs' own blog shows how tight that link is: in July, Fyxer, an AI inbox assistant, made ElevenLabs the exclusive transcription provider for its meeting notetaker. ElevenLabs says Fyxer's trial-to-paid conversion rose from 17.6 to 20.3 percent as a result. That is a vendor number, but the point stands: the model is a measurable part of the product.

And that part is now sitting in the shop window as a standalone app. My rule of thumb: your code is no longer a moat, your distribution is. If all you sell is a nice interface on top of someone else's model, you have a feature, not a business. The model vendor sees usage, costs and the weak spots of every app built on it. According to The Next Web, it also has roughly $600 million in revenue behind it.

The catch: note-taking is turning into a commodity

For buyers, this is good news. When the supplier gives away a note app, plain transcription becomes table stakes. The question shifts from who transcribes most accurately to what happens to the notes afterwards. Do they land in the CRM? In the inbox as a ready draft? Does a bot visibly join the call or not? That is the difference from the annoying status quo, and it is what you should choose a tool on, not the word error rate on a spec sheet.

For the note-taking apps, the way I see it: those with their own workflows, their own data and a brand people recognize in daily use will survive the switch. Those selling transcript plus summary are now in a direct price comparison with their own supplier.

What I would do now

• Before buying the next note-taking tool, ask which model is inside and what the tool does that the model vendor's own app does not.
• Test a real workflow instead of a demo: one internal meeting, one with a customer, then check where the tasks end up.
• If you build on someone else's model yourself, write down what your product is worth without it. If the list is short, that is your roadmap.

Bottom lineOnce your supplier rebuilds your app, you are selling the workflow, not the model.