Reimagining Knowledge Work with AI: Ivan Zhao Connects the Past to the Future at Notion

Notion CEO Ivan Zhao reveals the philosophical foundations behind his company's success, from democratizing computing tools to building the infrastructure for knowledge work AI agents.

Ivan Zhao shares his journey from China to Silicon Valley, the early computing philosophy that inspired Notion, and how the company is positioned to lead the transition from traditional SaaS tools to AI-powered knowledge work agents.

Key Takeaways

• Notion's strategy of consolidating multiple SaaS tools into one platform positions them uniquely to build "knowledge work agents" that require integrated context and tools
• The transition from SaaS to AI represents a shift from "selling tools to facilitate work to providing the work itself" through AI agents that can reason and act
• Building AI products requires a fundamentally different approach - more like "brewing beer" (organic, unpredictable) than "building bridges" (engineered, predictable)
• Early computing pioneers envisioned software as a medium everyone could shape, not just programmers - a vision that modern no-code tools and AI are finally realizing
• Craft and aesthetic attention to detail remains crucial even at scale, but requires balancing perfectionism with business utility (aiming for 7/10 rather than 10/10)
• The "agent" buzzword represents software that combines knowledge, skills, assignments, and context to perform actual work rather than just providing tools
• Physical environment design principles mirror software design - honest materials, lived-in feel, and avoiding excessive polish that suggests misplaced priorities

Timeline Overview

• 00:00–20:00 — Aesthetic Philosophy and Personal Style: Ivan's attention to craft details, from office wood selection to fashion choices, and how simplicity in some areas (food preferences) balances complexity in others
• 20:00–40:00 — Cultural Immersion and Learning: Moving from China to Canada at 17, self-educating on Western culture through systematic consumption of five-star music albums and Criterion Collection films
• 40:00–60:00 — Silicon Valley Journey and Computing Philosophy: Working at Inkling, discovering early computing pioneers' vision of democratized software creation, and the inspiration to build tools everyone can shape
• 60:00–80:00 — Early Notion Development: Years of product-market fit struggle, the Japan retreat for rebuilding, and the philosophical foundation of making software malleable for non-programmers
• 80:00–100:00 — AI Era Transformation: The Cancun hotel room sprint to build AI features, recognizing the shift from tools to automated work, and positioning Notion for knowledge work agents
• 100:00–END — Management Philosophy and Future Vision: Interview techniques, daily routines, company culture around craft, and the evolution toward AI-powered knowledge work automation

The Aesthetic Foundation: Craft as Competitive Advantage

Ivan's approach to aesthetics and craft reveals sophisticated thinking about how attention to detail creates both product differentiation and cultural coherence within organizations.

• Physical Environment as Culture Signal: Ivan's obsession with office details - from wood veneer quality to craftsman-style paneling - stems from belief that "when you make the room good, people feel it." The office serves as a recruitment tool where "candidates work from our office" and see "we do care so much about a lot of details, a lot of craft including our people."
• The Lived-In Philosophy: Rather than pursuing perfect polish, Ivan advocates for environments that feel "lived in" and "malleable" where "you should be able to move things around" because "if you have too good office your business go down because you're not focused on the right thing." This reflects deeper product philosophy about balancing craft with utility.
• Intuitive Quality Detection: Ivan describes his ability to sense software bugs as similar to aesthetic perception - "when things off a little bit I can sense it, I feel it." This extends to filing bugs about millisecond differences in typing lag that most users wouldn't consciously notice but would subconsciously experience.
• The Gap Between Taste and Ability: Ivan articulates the hierarchy of "your taste and ability to describe there's a gap and ability to make there's another gap." As a software maker, he can identify bugs, understand their technical causes, and know how to fix them - a complete loop from perception to solution.
• Strategic Craft Optimization: Notion deliberately targets "6.5, seven out of 10" quality rather than perfection because "if you push that too far, you're optimizing too much for the craft and maybe beauty, but not enough for the business and utility." This calculated approach prevents craft from becoming counterproductive to business objectives.
• Simplification Counterbalances: Ivan's preference for simple food ("burrito guy" rather than "five course Michelin style dinner") may represent necessary simplification in areas that aren't core competencies, allowing mental energy to focus on areas requiring deep attention to detail.

Computing Democracy: The Philosophical Foundation of Notion

Ivan's vision for Notion stems from deep study of early computing pioneers who intended software to be a moldable medium accessible to everyone, not just programmers.

• The Hippie Computing Origin: Modern computing emerged from "this group of people the hippie generation 60s 70s" who envisioned computers as interactive creative tools rather than room-sized calculators for "calculating taxes or ballistic missile trajectories." This generation invented the mouse and graphical interfaces to make computing more human-accessible.
• Software as Universal Medium: The original intent was that "computing are just like reading and writing. It's a medium that you can shape" where "some really good people can be a poet, some people can be a novelist or you can use for business work." This vision sees software as a creative medium with unlimited expressive potential.
• The Programmer Class Problem: Despite the original democratic vision, "until recently computing is by and large the medium that only the programmer class can shape" making "programmers kind of like the scribes of the modern era" where "only few people can read and write" in the digital medium.
• Steve Jobs' Incomplete Vision: While Jobs popularized personal computers by "putting PC on every desk," he also "locked computing, locked software into this application prison" that's "easier for people to understand in some ways but a lot more rigid" and "cap the ceiling of how far you can go with your software."
• Object-Oriented Computing Vision: The unrealized third innovation from Xerox PARC was "object or environment" design that would allow "computer to be like Lego-like object-like people can take apart put it back together without being programmers." This modularity represents the true democratization of software creation.
• Modern AI Democratization: Current AI tools are beginning to realize the original vision where "more people can vibe coding" through natural language, though Ivan cautions we're still in early stages of understanding how to use this new medium effectively.

The SaaS Consolidation Strategy: Building Infrastructure for AI Agents

Notion's business strategy of consolidating multiple productivity tools into one platform creates unique advantages for the emerging AI agent era.

• The Fragmentation Problem: Traditional knowledge work requires "a dozen different tasks across a dozen different apps" making it "extremely difficult to create knowledge work agent because knowledge work today are those fragmented" across disconnected systems.
• Context and Tools Integration: AI agents require "the context and tools to be together" because "language model want the context to be together so they can start thinking and problem solving for you." Fragmented systems make this integration nearly impossible for AI to navigate effectively.
• The Coding Agent Advantage: Programming agents succeeded first because "most for coding agents all the contexts are in the GitHub repos in one place, plain text file easy to read and write and the tools it needs to use are reading and writing those files so it's very self-contained."
• Notion's Unique Position: As "one of the few companies in the past five plus years actually been consolidating those contacts and tools," Notion has assembled the building blocks necessary for comprehensive knowledge work automation rather than point solutions.
• Enterprise Search Integration: Beyond internal consolidation, Notion's "enterprise search product integrates with 10 plus different external contacts from the Google stack to Microsoft stack, Atlassian stacks" enabling AI agents to work across organizational tool ecosystems.
• Compound Productivity Effects: The consolidation strategy delivered immediate SaaS-era value (RAMP "reduce their tooling cost by 70%" while shipping "things much faster") that now enables AI-era advantages where "each employee has those three four AI agents out of the box."

Beer vs Bridges: The New Paradigm for AI Product Development

Ivan's brewing metaphor captures the fundamental shift required in product development approaches when working with AI systems that have organic, unpredictable capabilities.

• Classic Software as Bridge Building: Traditional software development follows predictable engineering where "you sort of can engineer anything you can imagine" through either "Y Combinator style follow your customers" or "Steve Jobs style you have a vision you don't listen to customer" but both approaches can systematically build toward defined specifications.
• AI as Beer Brewing: Language models "work differently you cannot build everything you can imagine" because they "often time take you 70 80% there and never close that gap of remaining 20%." Like brewing, "you have to create the right environment, just channel the yeast to let them do their work."
• The Organic Control Problem: With AI, "you cannot tell the yeast, hey yeast, please taste that way. You cannot force it" requiring developers to "channel the model" rather than direct it. "The wisdom and capabilities is inside this model. Sometimes it can build the thing in your head. Sometimes it doesn't."
• Process Sequence Changes: AI development requires "experiment a lot. You put design engineer side by side together and with the data with the right context and just keep trying things and see what sticks" rather than the traditional waterfall approach of vision → specification → engineering.
• The Demo vs Product Gap: This explains why "there's so many cool demos, but now harder to find real product because to be a real product, you need to close the gap of remaining 20 30%" which may be impossible to force through traditional engineering approaches.
• Organizational Adaptation Challenges: The brewing approach "changes the sequence of how the software development process" and becomes "really hard for large companies once the company gets large hard to shift this way of building software" because it requires abandoning established predictable processes.

Knowledge Work Agents: The Next Computing Paradigm

Ivan positions AI agents not as enhanced tools but as a fundamental shift from providing capabilities to performing actual work, representing the biggest change since personal computing.

• From Tools to Work: The transformation involves "transitioning from selling tools to facilitate work to providing the work itself" where software companies can now provide "not only the tools but providing the human sitting behind the tool as a package" because language models have "human capability to reason to think to understand."
• The Human in the Machine: When Ivan says "there's a human in there," he means language models possess cognitive capabilities that enable them to perform knowledge work tasks rather than just assist with them. This represents a categorical shift from automation to artificial intelligence.
• Knowledge Work Precedent: Customer support agents already demonstrate this transition where systems now do "the support" rather than just making "your support agents more efficient." Coding agents similarly provide "final output" rather than just IDE assistance like "tab completion."
• Three Foundational AI Products: Notion's initial AI offerings show the pattern - enterprise search "generates answer for you" instead of requiring manual research, research agents "draft a report based on all your internal content" replacing hours of synthesis work, and meeting notes that are "better than any human can write."
• Context Window Importance: AI agents require comprehensive context to function effectively, which explains why "language model wants the context and tools to be together" and why fragmented SaaS systems create barriers to effective agent deployment.
• The Timing Advantage: Unlike previous AI winters, current capabilities feel different because "the underlying capability you know is just so powerful" and "the capability gets better every three-ish month" with no apparent stopping point, suggesting sustained progress rather than hype cycles.

Craft at Scale: Managing Quality Across Expanding Surface Area

Ivan's approach to maintaining craft standards while scaling reveals tensions between perfectionism and business pragmatism that every growing company must navigate.

• The Surface Area Challenge: High craft becomes exponentially harder with complexity - "it's easier to create a jewelries" like the German to-do app Things that "can only do like 10 features" where "everything feels great" compared to building "something much more complex like a notion, like a linear, which is much larger surface area."
• Entropy and Software Evolution: Ivan acknowledges that "entropy erodes everything including software" and points to Microsoft Word as an example where "there's layers of layers of functionality and people's needs and wants packed into that thing" making simplicity nearly impossible to maintain.
• The Reset Necessity: Without periodic resets, "it's really hard to build something that's simple, shrink the surface area, then still provide a utility and still provide high craft" because companies default to increasing features while dropping craft standards.
• Physical Tool Inspiration: Notion's conference rooms are named after enduring physical tools like "iPhones, the original BMW 3 series, Toshiba rice cookers, Sony transistor radio, Singer sewing machines" that "last decades sometimes 100 plus years" because "they really find something that really works for people."
• Form Factor Stability: Great tools achieve lasting form factors that "can last two decades, three decade, five decades sometimes" because they solve fundamental human needs with appropriate material constraints rather than chasing feature proliferation.
• Quality Calibration: Rather than pursuing perfection, Ivan targets specific quality levels knowing that "if you push that too far, you're optimizing too much for the craft and maybe beauty, but not enough for the business and utility" suggesting quality should be deliberately calibrated to business context.

Management Philosophy: Taking Off the Business Jacket

Ivan's interview and management approach reflects belief that authentic human connection produces better outcomes than formal professional personas.

• Interview as Conversation: Ivan's interview process focuses on "having a normal conversation" where "you need to take off their virtual business jacket" because fundamentally "it's a match" between candidate values/skills and company needs rather than an examination or judgment.
• Reference Over Interview: While formal interviews provide limited insight ("one hour or panel of four or five people that's only four or five hours is very like blind person touching elephant"), back-channel references from "people that spend years with the other candidate" provide much better evaluation data.
• Craft-Based Evaluation: For technical roles, Ivan "goes deep on the craft" because discussing specific work reveals both competency and values naturally, providing more authentic assessment than abstract questions about culture fit or personality traits.
• Non-Procedural Approach: Rather than following checklists, Ivan uses "a bag of different questions or tricks and you just follow your notes after that" because rigid procedures prevent genuine conversation and "you don't get the realness out of it."
• Human Recognition: The process starts with ensuring "the other person understand you are a human and you see them as a human" rather than maintaining artificial professional distance that prevents authentic evaluation of potential working relationships.
• Values Integration: Company values emerge naturally through craft discussions rather than requiring separate evaluation because "from talking about craft you get a sense their value too" in ways that feel organic rather than performative.

Daily Rhythms: Simplicity Enabling Complexity

Ivan's personal routines reveal how deliberate simplicity in some areas creates capacity for deep engagement in areas requiring intensive focus.

• Morning Writing Priority: Ivan's routine prioritizes creative work by waking at "6:30 7ish," making tea and coffee, then writing before checking communication channels because "I try not to let my Slack notification come to me" and "really important things will come to my text messages."
• Notification Discipline: By turning off Slack and email notifications and choosing when to engage with them, Ivan maintains control over attention and prevents reactive patterns that fragment deep work time.
• Evening Learning Ritual: Bedtime involves reading "philosophy or history" or exploring crafts like "Japanese woodcut master from early 1900s" through books and YouTube videos, using intellectual curiosity to wind down rather than mindless entertainment.
• Exercise Integration: Running "two and a half miles" fits into the morning routine after writing, providing physical activity that doesn't compete with peak mental energy periods but supports overall cognitive function.
• Sleep Optimization: Maintaining seven hours of sleep and an 11:30 bedtime creates consistent recovery cycles that support sustained intellectual work without requiring willpower to maintain basic self-care.
• Context Switching Management: The routine creates clear boundaries between creative work (morning writing), physical maintenance (exercise), and reactive work (communication) rather than mixing these different types of attention demands.

Common Questions

Q: How does Notion's consolidation strategy create advantages for AI development?
A: By bringing context and tools together in one platform, Notion enables AI agents that can work across knowledge work tasks rather than being limited to fragmented, single-purpose applications.

Q: What makes building AI products different from traditional software development?
A: AI development is more like "brewing beer" - organic and unpredictable - rather than "building bridges" where you can engineer anything you imagine through systematic specification and development.

Q: Why does craft and aesthetic attention matter for software companies?
A: Physical environment design reflects company values and attention to detail, serving as both recruitment tool and cultural reinforcement that extends to product quality.

Q: How should companies balance craft quality with business utility?
A: Target deliberate quality levels (like 7/10) rather than perfection, recognizing that excessive craft optimization can detract from business value and utility.

Q: What was the original vision for computing that inspired Notion?
A: Early computing pioneers envisioned software as a moldable medium accessible to everyone, not just programmers - similar to how writing enables poets, novelists, and business communication.

Ivan Zhao's vision for Notion represents a sophisticated synthesis of aesthetic philosophy, computing history, and business strategy. His approach combines deep attention to craft and detail with pragmatic recognition of business constraints, while positioning the company to lead the transition from traditional productivity tools to AI-powered knowledge work agents. The success of this vision depends on maintaining cultural coherence around quality and craft while navigating the complex challenges of scaling both technology and organization in an era of rapid AI advancement.

Practical Implications

• Consolidate tools and context within platforms to enable AI agents that can work across multiple knowledge work tasks rather than building point solutions

• Adopt "brewing" approaches to AI product development that emphasize experimentation and channeling AI capabilities rather than forcing predetermined specifications

• Balance craft quality at deliberate levels (like 7/10) that serve business utility rather than pursuing perfection that may detract from practical value

• Create physical environments that reflect company values and attention to detail as both cultural reinforcement and recruitment advantages

• Implement interview processes focused on authentic conversation and craft evaluation rather than formal procedures that prevent genuine assessment

• Establish daily routines that prioritize creative work before reactive communication and maintain clear boundaries between different types of attention demands

• Study computing history and philosophy to understand the democratic potential of software tools rather than accepting current limitations as permanent

• Position products for the transition from providing tools to providing actual work through AI agents that can reason and act autonomously

• Recognize that managing AI systems requires fundamentally different approaches than traditional software engineering and plan organizational changes accordingly

• Build platforms rather than applications when possible, since AI capabilities benefit from integrated context and tools rather than fragmented experiences

• Maintain human-centered design principles even as AI capabilities advance, ensuring technology serves human creativity rather than replacing it

• Develop aesthetic sensibility and attention to quality details that compound into competitive advantages over time