Notion's AI Pivot: Building the Future of Online Communities

From Docs to AI Workspace: Notion's Radical Reinvention

Notion, the productivity platform beloved by individuals and teams alike, is undergoing a metamorphosis. Launched in 2013, it grew from a simple note-taking app into a sprawling workspace for documents, databases, and project management. By 2021, it was valued at $10 billion. Fast forward to 2025, and Notion is worth around $11 billion, with annual recurring revenue north of $600 million—and half of that comes from AI products. This isn't just a story about adding a chatbot. It's a tale of how a successful SaaS company decided to essentially rebuild itself to avoid becoming obsolete.

The Kyoto Reset: First Rebuild

Notion's early years were rough. Founder Ivan Zhao and co-founder Simon Last struggled to find product-market fit. They were running out of money and felt a deep despair—not knowing what was wrong or where to turn. So they made a drastic choice: they laid off their team and moved to Kyoto, Japan. Why Kyoto? Tokyo apartments were too small. In Kyoto, they found larger, cheaper spaces, and by renting out their San Francisco office and home, they achieved positive cash flow for the first time.

In Kyoto, the duo entered a simple rhythm: code, eat, repeat. With no resources or extra people to coordinate, Zhao had to confront the core question: What problem does this tool solve? Why does it matter? He didn't shut down Notion and start anew—he was obsessed with the idea of a tool that enhances human capability, a passion from his college days. The Kyoto reset wasn't about pivoting; it was about stripping away the non-essential and returning to the fundamental mission of building a genuinely useful human tool. This ethos—flexibility, modularity, and user-driven creation—became Notion's DNA and is what makes it so beloved in online communities today.

The GPT-4 Awakening: Second Rebuild

The second rebuild came around 2023. By then, Notion was a mature company with hundreds of employees, a successful product, and a high valuation. The natural path would have been to keep adding features, hire more salespeople, and optimize pricing. But Ivan Zhao got early access to GPT-4, and it shattered his assumptions. GPT-3 had felt merely 'useful,' but GPT-4 signaled a fundamental shift in how knowledge work would be done. Notion had already launched an AI writing assistant two weeks before ChatGPT, which boosted revenue, but Zhao wanted more. He envisioned agents that could understand context, retrieve info, execute tasks, and drive workflows—not just generate text.

That vision proved challenging. Notion started experimenting with agents in late 2022, but for a year and a half, they struggled. The technology wasn't ready; their ambitions were too far ahead. Zhao later admitted they were 'living too close to the future.' This experience taught him a crucial lesson: building AI products is like brewing wine, not building a bridge. You can't command the yeast to produce a specific flavor; you have to experiment, see what the model can do, and then adapt your product around it. This shifted Notion's development process. Instead of a linear pipeline from product manager to designer to engineer, design, product, and engineering now had to collaborate early, directly around model capabilities and user feedback.

From Toolbox to AI Workspace: Serving Online Communities

Notion's core strength has always been its flexibility. Unlike rigid office suites, Notion lets users build their own systems—pages, databases, tasks, and wikis all in one place. That becomes even more powerful with AI. Because AI thrives on context, and Notion holds a treasure trove of that context: documents, meeting notes, project updates, and historical decisions. That's why Notion is repositioning itself as an 'AI workspace,' not just a note-taking tool. Features like Notion Agent, Enterprise Search, AI Meeting Notes, and Notion Mail are designed to help teams—and the online communities they build—work smarter.

Three directions guide this shift. First, finding knowledge: Enterprise Search tackles the problem of scattered info across Slack, Gmail, and Google Drive, providing answers with sources. Second, structuring communication: AI Meeting Notes extracts decisions and action items, turning meetings into actionable data. Third, automating repetitive work: Agents can create workflows, update databases, and generate reports, letting AI work within your system. For online communities, this means smoother collaboration, better knowledge sharing, and less busywork.

Organizing Like a Jazz Band, Not a Marching Band

Zhao describes Notion's ideal organization as a jazz band, not a marching band. He's not advocating for no hierarchy—he believes hierarchy is human nature. Instead, he wants flexibility and improvisation within structure. A marching band executes predetermined moves; a jazz band responds to each other in the moment. In the AI era, where change happens weekly, Notion needs that jazz-like agility. Product plans can't be locked in six months ahead. Financial planning matters, but product strategy must be fluid. This philosophy extends to how teams work: cross-functional, adaptive, and open to new roles.

Hiring for Taste and Initiative Over Experience

AI is leveling the playing field for basic skills. Writing and coding are becoming commoditized. So Notion now hires for two things: taste and initiative. Taste means knowing what's good and why. Initiative means acting in uncertainty. The company uses a 'barbell' structure: a few senior architects paired with many young, energetic engineers. Seniors provide direction and judgment; juniors use AI tools to execute. This trains the next generation while amplifying senior wisdom. Notion also values hybrid talents—designers who code, PMs who prototype. Even sales hires must build something with Notion and share a link, proving they can use tools, not just talk.

Marketing Split: Storytelling and Sales Enablement

Traditional marketing departments are too slow for AI-speed product changes. Notion broke its marketing into two tracks: storytelling (product narrative, content, community) and demand generation (leads, growth, sales support). The first sits close to product; the second close to revenue. This reflects Notion's community-driven roots—early users spread the word through YouTube and template sharing. But for enterprise sales, they realized they couldn't reinvent the wheel. Zhao admitted they'd erred in trying to bypass traditional sales with a magical growth hack. Enterprise buyers want human contact, especially for costly, sensitive software. So Notion focuses its innovation on product and AI workflows, not on reinventing sales.

The AI Transformation Isn't Just About Cutting Costs

Many companies tout AI for cost savings, but Notion sees bigger potential. AI does introduce real costs—each model call has a price, impacting margins. Zhao acknowledges they must be thoughtful about costs. Yet the transformation aims to change knowledge work itself. Historically, corporate knowledge management fails because information is scattered and hard to maintain. AI can search, summarize, and act on that knowledge, making it alive. This is Notion's edge: it already houses the work context for millions of teams. AI can now tap into that context, turning it into a strategic asset. For online communities, this means AI can help manage shared knowledge, moderate discussions, and surface insights, strengthening the community's collective intelligence.

Founders Must Feel AI Themselves

Zhao's advice to other SaaS founders: don't just read reports or watch videos. Get your hands dirty. Build something with AI, integrate it into your product, and see what paths open. Every company is different, and AI won't automatically tell you how to adapt. This is especially critical for established companies, which risk being locked into old processes. Notion's answer is clear: if starting today, it wouldn't be just a docs tool. It would be an AI-native workspace. That's what 'starting over' means—re-examining everything through new technological lenses.

Lessons for Online Communities and SaaS Companies

Notion's journey offers concrete lessons. First, AI transformation can't stop at adding a chat window. It must change how work gets done. Second, embrace uncertainty—experiment, evaluate, iterate. Third, don't blindly follow the SaaS playbook. Roles blur, and value shifts to those who understand users, tech, and can make judgment calls. Fourth, don't dismiss enterprise sales. Trust is still human. Fifth, founders must be hands-on, not just delegating. Finally, old companies can beat new ones if they're willing to rebuild. The hardest question for many will be: Are you brave enough to use AI to reimagine your company—and your community?