This week I logged onto my computer, joined a meeting, and got a pop up: Wispr Flow can now transcribe my meetings. It’s the Mac app I love for quick dictation, and meeting transcription is a logical extension of a product built on converting voice to text. But it got me thinking about how SaaS products go through waves of converging on the same feature.

It’s not just Wispr Flow either. Join a meeting now and I get a stack of pop ups from Notion, Fireflies, Granola, Gemini and Wispr Flow, all fighting to transcribe the same conversation.

The convergence cycle

Note takers existed long before the AI wave, but Granola set the trend for the lightweight AI notetaker when it launched in 2024. Notion joined in 2025, adding AI Meeting Notes to its calendar. The big platforms are all in too: Zoom’s AI Companion summarises meetings, Google Meet takes notes for you with Gemini, and Microsoft Teams has intelligent recap. Now Wispr Flow is in the mix as well. The feature everyone is currently converging on is note-taking.

We’ve seen this cycle before.

Appointment links went through it. Calendly proved the demand, and by the early 2020s scheduling was bundled into the CRMs, Google Calendar itself, and Outlook with Bookings.

Around 2022, it was whiteboards. Miro proved people wanted a shared canvas, Zoom bolted one directly into meetings, Figma launched FigJam, Canva added whiteboards, and even Apple joined the party with Freeform.

Async video followed the same arc. Loom made the quick screen recording a legitimate way to communicate, then Slack added Clips, Zoom shipped its own Clips, Dropbox launched Capture, and Vimeo built Record.

The pattern is consistent: a focused product proves demand for a feature, and within a couple of years it’s a checkbox in every adjacent suite.

Why this wave is different

In previous cycles, bundling a feature was about keeping users inside your product. This time the context has changed. We’re using AI tools to shuffle information between products, and the AI tool is becoming the operating system. I’d rather have the best operating system and connect it to the best in class tool for each job, not settle for whichever suite happens to bundle everything.

That changes what “best in class” means.

Why I switched notetakers

For a couple of years I used Granola and loved it. The note-taking experience is still industry leading. But I recently moved to Fireflies, and the reason had nothing to do with the base technology of transcription and notes.

What made me switch is what happens after the meeting. Everything ends up in an LLM eventually, so the thing that matters most is being able to retrieve the information cleanly. At StackOne, building MCP servers is our product, and we’ve found that an MCP built on top of a provider’s API can be more powerful than the first party MCP itself: you can shape the tools around your workflows rather than taking whatever the vendor exposes. Either way, it only works as well as the API behind it. A great note-taking UI matters less when the primary consumer of your meetings is an agent rather than a human scrolling back through notes.

The second thing is accuracy and attribution. Are speakers correctly attributed? Does the transcription pick up our terminology (and understand my Aussie slang)? These meetings are becoming the foundation for creating tickets, generating content, and writing feature pages. When the transcript is the source of truth, every error compounds through everything built on top of it.

The foundational layer wins

Note-taking becoming a commodity is good news. The baseline keeps rising, and the tools that thrive will be the ones that lean into being part of a larger system: capturing your meetings accurately and making it effortless for the rest of your stack, human or agent, to build on top of them.

Even as the software on top becomes less deterministic, a good API matters more now than it ever has.