The Great AI Rebundling
How AI turned once-distinct companies into direct competitors overnight—and why Figma now battles PowerPoint, Canva, Webflow, Miro, Illustrator, and Cursor all at once.
In 1995, Jim Barksdale was wrapping up Netscape's IPO roadshow when a banker asked the question on everyone’s mind: "How do you know that Microsoft isn't just going to bundle a browser into their product?" Jim responded:
"Gentlemen, there's only two ways I know of to make money: bundling and unbundling."
With that single sentence, he captured the rhythm of the tech industry perfectly. For three decades, we've watched this cycle play out: incumbents expand by bundling more products into their ecosystems, startups break through by unbundling with sharper, more focused products.
But right now, we're witnessing something unprecedented: the fastest, most aggressive rebundling wave in tech history.
I recently joined
and on the Unsolicited Feedback podcast to discuss why everyone seems to be building everything—at ludicrous speed.In just the past month, Notion launched an AI-powered email client, AI meeting notes, enterprise search, and research capabilities—essentially becoming the office competitor they always hinted at being. Figma debuted AI-powered tools for creating sites, app prototypes, and marketing assets, taking on Canva, Adobe, WordPress, and even coding tools like Cursor. Loom launched a meeting recorder to compete with Granola. Anthropic added integrations and search. OpenAI acquired Windsurf and launched Deep Research.
To understand what's happening, let's look at how we got here. Both Notion and Figma followed the classic unbundling playbook before their recent rebundling spree.
Unbundle to disrupt, bundle to expand
Notion, from teams documents to office suite
Notion began with a simple observation: Google and Microsoft had created document tools for the old world—a world of files managed by individuals. But as teams moved their entire workflow to the cloud, they needed something different. They needed a way to create, organize, and share an ever-expanding universe of team documents.
This is exactly how unbundling starts. Bundles, in their quest to serve everyone, inevitably serve no one particularly well. The cracks appear. And startups slip through.
Notion started by unbundling. But now they've launched email, meeting notes, enterprise search, and research tools—essentially rebuilding the entire office suite they originally disrupted.
What's remarkable is the speed shift. Notion historically moved with careful polish, launching one thoughtful product at a time. Calendar was a big deal. Now they're launching six major products simultaneously, each competing directly with established players in their respective spaces.
Figma, from UX design tool to creative suite
Figma spotted a similar opportunity. Adobe's Creative Suite had been the design standard for decades, but it was built for a world where designers worked in isolation, passing static files back and forth like notes in class. And Adobe had never really focused on the specific needs of UX designers.
Figma unbundled UX design from Adobe's creative empire. For years, they perfected this single vision—making design collaborative, accessible, and fast. But in recent weeks, they've launched tools for creating websites, generating code, bulk asset creation, and vector editing—moving far beyond their design tool origins into development, marketing, and creative workflows.
They are building a new type of creative suite, organized around the collaborative workflows that are essential to modern design.
Fueled by AI, Figma and Notion have moved from "we do one thing really well" to "we're building the end-to-end platform for your workflow."
How AI enables bundling
Figma and Notion are not alone. We’re seeing companies, across the board, build and expand their bundles. AI native products, like ChatGPT and Claude, are already bundles with multiple surfaces built-in for document creation, code editing, charting, spreadsheets, and more.
AI is driving bundling for three reasons:
AI is the most general-purpose technology ever created
AI products benefit from being integrated into the workflow
AI has accelerated product development
Let’s take a closer look.
AI is the most general-purpose technology ever created
In the past, software was necessarily specialized. It would have been prohibitively complex to manage a single codebase that handled documents, code, spreadsheets, images, and audio all at once. AI changes this fundamental constraint. The same foundational models that power writing can handle analysis, coding, design, and communication. Companies can now build sophisticated features across completely different domains using the same underlying technology.
Figma's new tools exemplify this shift. They rely on Anthropic's Claude Sonnet AI model, enabling them to generate everything from marketing copy to functional code. The technical barriers to expanding product surface area have never been lower.
AI products benefit from being integrated into the workflow
But there's a deeper strategic reason driving this bundling wave: in the AI era, workflows are king. As we discussed in the new Reforge AI Strategy program, being central to a user's workflow has become the new competitive moat.
The companies winning the rebundling race understand a critical insight: being a point tool is increasingly dangerous. If you're adjacent to the workflow rather than central to it, you're vulnerable to being absorbed by larger platforms that already own that workflow. Jasper learned this the hard way as tools like Notion added AI-powered writing, eroding their foothold as a standalone writing solution.
This workflow-first thinking explains some recent moves in the market. Grammarly saw the writing on the wall. Their acquisition of Coda wasn't about expanding features—it was about owning the workflow. They realized that just correcting text wasn't enough; they needed to embed into where people actually do their work. The companies expanding most aggressively today are those doubling down on the customer workflows they already own.
AI has accelerated product development
For decades, companies dreamed bigger than they could build. Creating a comprehensive bundle required massive engineering resources, extended timelines, and enormous risk. Most companies had to choose: focus on your core product or spread yourself thin trying to compete everywhere.
AI eliminates this trade-off. Companies are no longer rate limited by what their engineering teams can deliver. We've heard from multiple sources that companies are now producing software faster than they can release it. This explains the breathtaking pace of new releases.
Now, companies can close the gap between their strategic aspirations and the products they have in market. Notion always wanted to be an office suite competitor, but building email, search, and meeting tools would have taken years in the past. Now they can launch all of them—all at once.
But here's the catch: while companies can move at AI speed, people still move at human speed. This creates a fundamental tension in the rebundling strategy.
The adoption challenge: building faster than users can absorb
Personally, I’m an active Notion user. Yet, I didn’t find out about Notion’s email client until we started talking about it for the Unsolicited Feedback podcast (more than a month after it launched).
Major product launches are getting lost in the noise—a warning sign for every company pursuing this strategy.
The discovery problem reveals a fundamental flaw in the rebundling strategy. Companies are adding features faster than users can adopt them. Users end up employing only a sliver of what's available, which defeats the entire purpose of bundling. The weak link has shifted, as Brian noted on our podcast:
"The constraint has moved from how much you can ship to how fast the market and your users can absorb changes."
This creates an ironic opportunity: the same AI driving rapid product development might also solve the adoption crisis it created. Fareed posted a key question:
"As products get wider and wider and wider and do more stuff, how are people going to use AI more effectively to help onboard, help customers use and adopt all these various capabilities?"
When to bundle, when to stay focused
So how should companies navigate this new landscape? The answer depends on where you sit.
Incumbents → bundle to expand
If you already own central workflow real estate, speed is everything. The rebundling opportunity is massive, but it won't last forever. Ask yourself:
Workflow centrality: Are you where work begins, or just where it ends up?
Speed imperative: Can you expand before point solutions establish footholds in areas we should own?
Data advantage: What unique context enables you to deliver better AI experiences than competitors?
The most vulnerable players are those with slow iteration cycles. Notion moved aggressively to weave AI throughout their platform. Google Workspace and Microsoft Office added AI features too, but they feel bolted-on by comparison—sidebar copilots rather than reimagined experiences.
Bundling works particularly well with enterprises, who prefer working with fewer vendors they can verify meet security and compliance requirements.
Startups → find your wedge
The startup opportunity isn't dead, but the game has changed dramatically. Success requires:
Customer-first approach: People's problems haven't changed, but available solutions have transformed fundamentally. Consider Stack Overflow's rapid decline as developers flocked to GitHub Copilot and Cursor. Same problem—getting unstuck on challenging code—but AI assistants solve it radically better than StackOverflow’s Q&A forum. Breaking through as a startup means finding that unmet need or delivering a fundamentally better solution.
Differentiation test: For the first time, startups across every domain compete against a single product: ChatGPT. There's never been a tool equally capable at therapy and data analysis. The critical question: Can you create something 5-10x better than ChatGPT for your specific use case? If not, you'll struggle to break through.
Flywheel foundation: What unique data advantage can you build that creates compounding returns? Volume matters, but marginal value matters more—specifically, the value of your data relative to what's already in training sets. Real-time information like news, prices, and inventory is incredibly valuable because it exists outside of AI training datasets—models simply can't access live, changing data without special integrations. Meanwhile, informational and editorial content—the kind that powered SEO businesses like Stack Overflow or WebMD—has lost its moat entirely.
Unbundling works particularly well for individuals and SMBs who prefer simpler, more affordable, purpose-built solutions.
For the first time, startups across every domain compete against a single product: ChatGPT. There's never been a tool equally capable at therapy and data analysis.
Everything, everywhere, all at once
Jim Barksdale was right all those years ago. Markets sway back and forth—as one company expands by bundling, another finds a wedge by unbundling.
Today, this continues to be true, but the pace has fundamentally changed. Figma's recent launches now pit them against PowerPoint, Canva, Webflow, Miro, Illustrator, and Cursor—all at once. What were once distinct market segments are blurring together at unprecedented speed.
As we discuss in the Reforge AI Strategy program, AI is not like prior tech shifts. It’s creating more Red Ocean than Blue Ocean, and the bundle brawl is one more reason why today’s strategic environment is so intense.
The implications are clear. Incumbents must move quickly to own their workflows end-to-end, vigilant that once-unrelated competitors don't chip away at their position. Disruptors need to find their wedge by working customer-first and identifying opportunities where AI delivers 10x improvements.
The bundling and unbundling cycle continues, but AI has accelerated it to unprecedented speed. In this new reality, every company must choose: bundle aggressively to own workflows, find a disruptive wedge, or risk irrelevance.
Very deep pattern explanation! I really loved the “workflows ownership” concept.
But the problem of the adoption bandwidth stands out to me as an individual user. I think I have an AI fatigue😂 And trust issues, too (i.e. don’t trust AI to produce a high-quality output without time-consuming input). So where do I get enough time, mental capacity, and, frankly, patience to adopt new tools and features (which are often quite pricy, too)…
It was an interesting read!
The fact that organisations wanting one vendor and the nudge that existing incumbents can get ignored if they don't own the workflow really stuck with me!
Thanks for sharing, Ravi..
Also what do you think about Google's recent announcement of "checkout inside search"? This in short, will help people buy things without even getting to the products landing page/ checkout pages effectively..no longer people are gonna see the reviews or the 10% off things from one site..they'll get to see the best offer then n there.
Is this a way of them owning search to buy experience something that was kind of done in sites like amazon n others?