Notion CEO: AI Needs Human 'Taste' and 'Agency' by 2026
Serge Bulaev
Notion CEO Ivan Zhao says that AI can't replace two important human abilities: taste and agency. As more companies use AI to do regular tasks, people will stand out by using good judgment and making smart decisions. Instead of just doing work, workers need to choose the right goals and decide what feels right. Companies are now teaching employees how to review and improve AI's work, not just create it. This shift means jobs will focus more on human judgment and creativity, not just following rules.

Notion CEO Ivan Zhao's declaration that AI needs human 'taste' and 'agency' signals a major workforce shift expected by 2026. He argues that as AI automates routine tasks, professionals will be valued not for execution, but for their superior judgment and strategic initiative. This pivot redefines work, placing a premium on human-centric skills like curating AI output and setting meaningful goals, rather than simply completing tasks.
This isn't just a sound bite; it's a strategic forecast. Zhao believes these uniquely human qualities will determine career success as autonomous software agents become standard. With Gartner projecting that 40% of enterprise apps will feature such agents by 2026, the imperative is clear: workers must pivot from task execution to high-level judgment, prioritization, and decisive action.
Agency beats execution by 2026
Agency is the capacity to set goals, take initiative, and direct AI agents effectively, transforming human workers into strategic overseers. Taste is the critical judgment used to evaluate AI output for quality, brand alignment, and user empathy, ensuring the final product is not just complete, but refined.
In his essay, "Steam, Steel, and Infinite Minds," Zhao frames AI not as a tool for speed, but as a force multiplier for human effort. At Notion, this is already in practice: a co-founder who was once a 10x programmer now directs AI coding partners to achieve a 30-40x output, according to a Notion analysis. This model is being replicated across departments like finance and support, with Notion deploying over 700 internal AI agents.
As rote execution becomes a commodity, leadership focus is shifting to agency - the ability to define objectives, select the right AI tools, and validate their output. The payoff is significant: companies that pioneer these 'agentic' workflows are seeing 88% productivity boosts and halving their hiring cycles, based on recent talent data.
Taste surfaces as the scarce differentiator
While agency focuses on if a task can be done, taste determines how it should be done. Zhao defines taste as a blend of aesthetic sense, pattern recognition, and deep user empathy. An AI can generate ten versions of a landing page instantly, but only a human can decide which one builds trust or sparks delight. Consequently, forward-thinking companies are training employees to become expert critics of AI work, not just creators.
A quick scan of emerging programs shows the shift:
- Vodafone trains marketers to critique and reject generic AI copy during onboarding.
- Eaton grew its talent network by 300% by shifting interviews from keywords to narrative reasoning.
- MentorcliQ uses "taste mentors" to live-review AI research with junior staff, improving accuracy.
Hiring and upskilling pivot around human judgment
The hiring landscape is already transforming. AI screening bots now manage initial interviews, liberating recruiters to conduct deep, open-ended conversations that probe for initiative and resilience. Underscoring this trend, a World Economic Forum survey reveals that two-thirds of companies will prioritize AI literacy in 2025-2026 hiring, with 40% planning to reduce roles based on linear, repetitive tasks.
Corporate training budgets reflect this new reality. At Accenture, career progression is now linked to completing 'taste workshops' for critiquing AI-generated content. Similarly, Walmart's collaboration with OpenAI provides sandboxes where employees can practice prompting and debate the ethical implications of AI use.
Practical playbook for leaders
- Audit Workflows: Identify and delegate all rule-based, repetitive tasks to AI agents.
- Define Taste: Create clear rubrics to evaluate AI output on clarity, originality, brand alignment, and impact.
- Model Agency: Implement mentorship programs where senior leaders guide others in framing goals, crafting prompts, and validating AI results.
- Reinvest Savings: Track time saved by AI automation and reinvest a minimum of 30% into exploratory projects that cultivate human taste.
Implementing these strategies doesn't require massive R&D spending, but it directly supports Zhao's core thesis. As AI agents proliferate, competitive advantage will belong to companies that cultivate employees who possess agency - the drive to act decisively - and taste - the wisdom to know why a choice is right. This emerging contest for human judgment is already reshaping job descriptions, career paths, and corporate profitability.