Why Everyone Is Building AI Slides and Notetakers

August 19, 2026

Literally everyone is building the same two things: AI note-takers, and AI presentations.

I just spent 10 days in Japan meeting companies. The system integrators I met were building dedicated productivity software for companies. Almost every project began with capturing meetings and generating slides.

I met the new consulting arm in the Toyota group. They're working with a handpicked startup on new products. You guessed it: AI notetaker and slides.

I met Felo AI, a fast-growing consumer AI product. Their new feature? Slides. "We're taking customers from Gamma" the person proudly said.

Then I looked beyond Japan and realized everyone else was doing it too.

You already know that OpenAI, Anthropic, Google and the other foundation-model labs can all generate designed presentations.

General-purpose agents and research products such as Manus, Genspark, Felo, Perplexity and Kimi are turning research into decks too.

While I was in Japan, Napkin AI sent a broadcast email to users on their new product: slides. And Wispr Flow did so too. New product: AI meeting notetaker.

Then there are the dedicated products for both: Gamma, Prezi, Beautiful.ai, Pitch, Decktopus, SlidesAI, Chronicle and many more.

And of course dedicated meeting notetakers: Fireflies, Granola, Otter, Fathom, and hardware ones like Plaud and Pocket, with more coming out all the time.

This should be terrible news if you're a company building these products.

If every AI assistant can make a beautiful deck, why would anyone need a dedicated presentation product? The capability is becoming table stakes. Surely the standalone companies should be getting crushed.

But that is not quite what is happening.

Gamma is growing extraordinarily fast. Beautiful.ai recently announced a $45 million investment and says it has more than 100,000 business customers. Prezi is not a breakout growth company anymore, but it still claims 160 million users and maintains a team of more than 200 people. Pitch has downsized substantially, yet appears to have stabilized and continues shipping. Even across a crowded field with overlapping capabilities, many incumbents are holding their ground.

As for AI notetaker and meeting assistants, I happen to have talked to the investors of a few of these, and they universally told me the same thing: these companies are all thriving. Wispr Flow just raised another monster round.

How's that possible?

How can multiple of these companies be doing well when everyone from Google to Toyota to Accenture to no-name startups are doing exactly the same things?

A popular way to understand productivity startups was "The Unbundling of Microsoft". Microsoft Office contained Word, Excel, PowerPoint, Outlook and other tools. Startups could pull individual jobs out of that bundle and turn each into a better standalone product.

It was a useful frame. But if you imagine breaking a big market down into small individual markets, it would be a mistake

There were a certain number of documents to write, meetings to record and presentations to make. New software would compete to capture those existing jobs. Notion could take work from Word. Pitch could take work from PowerPoint. Slack could take work from Outlook.

Generative AI changes the equation because it does not merely move the work between products. It collapses the marginal cost of doing the work.

When the cost collapses, demand does not stay fixed.

It explodes.

Imagine that twenty years ago, someone offered your company an army of intelligent assistants.

They could attend every meeting, listen to every important conversation, capture what happened, distinguish decisions from random discussion, connect each conversation to the right project, and remember all of it forever.

Would you have limited them to the ten most important meetings each week? Of course not. You would have sent them everywhere.

The only reason organizations did not capture every conversation was that doing so would be prohibitively expensive.

AI note-taking changes that. Every meeting can become input to an organizational brain. Every conversation can become searchable context. An intelligent system can read, absorb, organize and analyze more of it than any human team could.

This is why every enterprise AI deployment seems to begin with note-taking. It is not because meeting transcription is the most exciting market in software. It is because every productive organization wants more memory.

Now imagine the same army of assistants could turn everything your organization knows into beautiful, audience-specific presentations.

One version for the board. One for employees. One for investors. One for a specific customer. One for a partner in Japan, translated and adapted to local context. A visual briefing before every important meeting. A tailored follow-up afterward.

Would you reserve them only for the quarterly board deck?

Again, of course not.

Historically, presentations were expensive. Even a short deck could consume hours of writing, layout and revision. So we saved them for occasions important enough to justify the labor.

I spent five years as a consultant making decks for companies. And I charged extra for every additional version: 8 slide teaser, 20 slide live version, 35 slide send-after… and of course I would charge extra, because each version took additional manual work.

That's all changed. The cost of producing any draft, any iteration, any additional version is now collapsing.

An investor asked for the deck of one of my portfolio companies that's raising money, I don't quite like the version they have because it's not tailored to this specific investor that I consider important. In the old days, I would ask my portfolio company to make a new version, they may get back to me in five hours, or in five days, or never. This time? I modified their existing version with one prompt, reviewed it, and sent it to the investor. All within 15 minutes, and done from my phone.

In this new world, the market is not "people who currently make PowerPoints."

The market is every occasion on which one person wants another person, or an entire group, to understand the same thing.

That is a much larger market.

Note-taking and presentations are the two ends of the same system.

Notes pull reality into the organizational brain. Presentations push organized meaning back out.

The better the input layer becomes, the more material there is to synthesize. The cheaper the output layer becomes, the more contexts become worth communicating. Each side creates demand for the other.

"Infinite" is obviously not literal. Human attention remains finite. But compared with the historical supply of people available to capture, process and communicate organizational knowledge, the latent demand is enormous.

What looks like dozens of companies fighting over one old software category may actually be dozens of companies entering a rapidly expanding flow of work.

If demand were fixed, ubiquitous AI generation should be brutal for dedicated presentation companies. Every deck made by ChatGPT would be one fewer deck made in Gamma, Prezi or Beautiful.ai.

But if the number of decks expands by ten or one hundred times, the arithmetic changes. A dedicated product does not need to preserve its share of the old market. It needs to capture a useful slice of a much larger one.

The foundation-model labs can make presentations. So can research agents, consultants, system integrators, office suites and vertical software products. That does not necessarily eliminate the dedicated tools. It can also train millions more people and organizations to expect presentations on demand.

And once people generate more presentations, they encounter all the work surrounding generation: editing, brand consistency, collaboration, permissions, source grounding, version history, analytics, delivery and reuse. The generic capability creates more demand for the specialized workflow around it.

This does not mean every presentation startup will win. Some are already shrinking. Generic text-to-deck generation will become free and undifferentiated. There will be consolidation.

But the activity itself is not disappearing.

It is becoming ubiquitous.

Then you come to the realization…

I am increasingly unsure that "note-taking software" and "presentation software" are markets in the old sense.

They may be closer to universal behaviors that every productive person, every organization and eventually every software product will need.

Every organization needs to absorb what is happening. Every organization needs to communicate what it knows.

Input and output.

Memory and expression.

Notes and presentations.

Generative AI did not invent those needs. It removed the labor constraint that forced us to satisfy them selectively.

That is why everyone is building AI slides, and everyone is building AI notetakers. It is also why at the very same time, the old slide companies and meeting capture companies are not simply disappearing, they're doing… fine.

The category is not being disrupted away.