Zeb Evans doesn’t check email anymore. He doesn’t scroll Slack threads. He doesn’t open project dashboards. Instead, an AI agent scrapes all of it on his behalf, decides what he needs to know, formats the essentials to look like a newspaper, and chats it to him each morning. “Most of my sidebar here is chats with agents,” the ClickUp CEO told Fortune in May 2026, giving a tour of his digital workspace. “I’ve got a few chats with humans, but the vast majority of them are with agents.”
Three days after that Fortune interview published, Evans cut 290 people from ClickUp’s 1,300-person workforce. Then he announced salary bands reaching $1 million in cash for whoever stayed and produced what he calls “100x impact.” The internet had questions.
The 100x org is a workforce model coined by ClickUp CEO Zeb Evans in May 2026, where AI agents outnumber human employees roughly 3:1, remaining humans earn salary bands up to $1 million per year, and the company restructures around three role types: builders, system managers, and front-liners.
It’s the most explicit version of something every SaaS founder is quietly considering: what happens when your AI agents can actually do the work?
Last updated: June 2026
Quick answers
What is ClickUp’s 100x org?
The 100x org is ClickUp’s workforce model where AI agents outnumber employees 3:1, the company is organized around three role types (builders, system managers, and front-liners), and surviving employees can earn up to $1 million per year for producing “100x impact” through AI system creation and management.
How many AI agents does ClickUp use?
ClickUp runs approximately 3,000 internal AI agents embedded across engineering, marketing, customer support, and product workflows. The company maintains an “agent org chart” listing every agent by name, its owner, and its per-run cost. CEO Zeb Evans told Fortune his Slack sidebar is mostly conversations with agents, not people.
Why did ClickUp lay off 22% of its workforce?
ClickUp cut 290 employees on May 21, 2026, framing it as structural rather than financial. CEO Zeb Evans said AI agents had absorbed the “structured, high-volume work” those roles handled, and the company needed to rebuild its org around agent-human collaboration rather than iterate on the old structure.
What is ClickUp’s 100x org?
ClickUp’s 100x org is a company structure built on one premise: AI agents have changed what it takes to build software, and the old org chart doesn’t map to the new work. Evans posted the announcement on X to his 229,600 followers on May 21, 2026. The core argument was blunt. Incremental improvements wouldn’t get ClickUp where it needed to go. The company needed to rebuild, not iterate.
The “100x” refers to a per-employee impact multiplier. It’s not a headcount target or a revenue figure. Evans is betting that a smaller team, directing thousands of AI agents, can produce output that a traditional team 10 times its size couldn’t match. The model has three structural components: a 3:1 agent-to-employee ratio, salary bands reaching seven figures for high-impact contributors, and a taxonomy of three surviving role types.
ClickUp is a $4 billion productivity platform based in San Diego. The company raised a $400 million Series C from Andreessen Horowitz and Tiger Global in October 2021 and hasn’t raised since. Revenue reportedly reached $278.5 million in 2024, with estimates projecting $500 million in 2026. The 100x org is Evans’ bet that the path to those numbers runs through fewer humans and more agents. At a time when Cognition AI just raised $1 billion at a $26 billion valuation for an AI coding agent, the market is clearly pricing in a future where agent-driven companies win.
How 3,000 AI agents actually work inside ClickUp
ClickUp went all-in on agents in January 2026, and by May the company had roughly 3,000 internal AI agents running across Slack channels, code review, customer support, marketing, and product workflows. That’s the highest publicly disclosed agent-to-employee ratio of any major SaaS company, per Fortune’s reporting.
The agents aren’t chatbots. They operate through a reasoning loop: observe context, decompose a goal into steps, select the right tools, execute, verify output, and escalate when uncertain. Each agent has its own memory systems, including episodic memory (what happened in past interactions), preference memory (how its owner likes things done), and both short-term and long-term storage. Fortune’s Sage Lazzaro described it as “a fundamental change to what work looks like for each employee.”
ClickUp tracks all of this through an agent org chart. Every agent has a name, an owner, and a visible cost-per-run. Evans discovered during a dashboard walkthrough that one agent costs $9 every time it fires. “Something to keep an eye on,” he said, “but not an immediate concern because it doesn’t run too often.”
Guardrails exist. No agent can delete data or merge code to production. When an agent needs information it can’t access, it messages its human owner directly. The control layer matters because AI agents at scale amplify errors without it.
Here’s what shifted in practice: employees used to spend their days drafting documents, evaluating data, reading emails. Now agents carry out that work. The human job is directing agents on what to do, reviewing what they produce, and giving feedback. Arianna Young, ClickUp’s principal of demand marketing, built an agent named Wall-E that coordinates speakers, scheduling, and execution for the company’s webinar program. She went from one webinar per month to six. “It’s able to do a lot of that manual task creation and coordination that used to be on my plate,” she told Fortune.
The learning curve is real. Wall-E once treated confirming a webinar moderator like a red-alert emergency, tagging in teammates because Young had described the task as “important.” She had to teach it that human hyperbole doesn’t translate well to agent instructions. “Being able to learn how to speak to the agents in a way that is extremely clear and doesn’t have some of that natural human hyperbole is a little bit of a different shift,” she said.
The three roles that survive in a 100x org
Evans described ClickUp’s future workforce in three groups. Each one has a different relationship with AI agents.
Builders are what Evans calls 10x engineers and 10x product managers. His claim: “The best engineers are not writing code any more. They are directing agents that write code.” The builders’ value comes from judgment, the ability to orchestrate agent output and catch what’s wrong. This tracks with what Andrej Karpathy predicted when he said AI agents would handle 99% of code generation by end of 2026.
System managers operate the automated workflows. They’re responsible for making sure the agent fleet runs as intended, monitoring costs, and handling escalations. Think of them as the people who keep the 3,000-agent fleet from going off the rails. The agent org chart runs through them.
Front-liners handle the work where human contact is irreplaceable: customer relationships, high-stakes negotiations, anything where empathy and real-time judgment matter more than speed. These roles aren’t shrinking because AI can’t do them. They’re staying because customers still want a person on the other end of certain conversations.
The three-role taxonomy isn’t just an internal memo. It’s a compensation structure. Evans said the path to $1 million salary bands is “available to nearly anyone in the company who produces 100x impact by creating or managing AI systems.” ClickUp has issued monthly Performance Awards including equity grants since 2022, averaging about five per month. In the past year, the criteria shifted to explicitly prioritize employees who use AI to drive outsized business outcomes.
Why did ClickUp lay off 22% of its workforce?
Evans framed the 290-person cut as architectural, not financial. The official line: ClickUp’s 3,000 AI agents had already absorbed the structured, high-volume work those roles handled. Keeping people in seats built for pre-agent workflows didn’t make sense. It’s the same logic that drove Replit’s CEO to bet the entire company on AI agents after revenue stalled at $2M ARR, though Evans had a much bigger ship to turn.
The timing raises questions. ClickUp’s last funding round was the $400 million Series C in October 2021 at a $4 billion valuation. That was peak-froth venture pricing. The company hasn’t raised since. Revenue has grown, but the gap between a 2021 valuation and 2026 economics is the kind of gap that makes “restructuring for AI” sound a lot like “restructuring for runway.”
Evans pushed back on that reading. He said savings from the layoffs would flow directly back into higher compensation for remaining employees. That’s where the $1 million salary bands come in. Whether this is a genuine structural transformation or PR packaging for cost cuts depends on whether ClickUp’s per-employee output actually increases in the next 12 months. The data doesn’t exist yet.
The broader tech industry has shed more than 100,000 jobs across roughly 250 layoff events in 2026 so far, according to layoff trackers. ClickUp isn’t alone in cutting heads and citing AI. The question is whether the 100x framing is a real operational playbook or a convenient narrative. As we covered in our breakdown of how to make money with vibe coding, the builders who thrive in 2026 are the ones who can direct AI tools, not the ones writing every line by hand. Evans is betting the same logic applies at the org level.
What other companies are trying this
ClickUp is the most public case, but it’s not the only one. Shopify CEO Tobi Lutke sent an internal memo in April 2025 telling managers they’d need to prove a job couldn’t be done by AI before requesting new headcount. Shopify’s employee count dropped from 11,600 in 2022 to 8,100 by end of 2024 while revenue grew at least 21% annually. Eight months later, much of the industry adopted the same hiring filter.
Klarna’s case is the cautionary tale. The Swedish fintech announced in early 2024 that its AI assistant was doing the equivalent work of 700 customer service agents, saving $40 million per year. Resolution time dropped from 11 minutes to under 2 minutes. By mid-2025, CEO Sebastian Siemiatkowski was telling Bloomberg the strategy had gone too far. CSAT scores dropped on complex tickets. Hallucinations surfaced on edge cases. Klarna started rehiring humans for what Siemiatkowski called “a VIP thing.”
The Gartner data backs up both sides. A survey of 350 executives found 80% of companies deploying autonomous AI have cut headcount. But workforce reduction rates were nearly equal among companies reporting higher ROI and those reporting modest gains or negative outcomes. Cutting people doesn’t automatically mean the math works.
A separate Gartner prediction estimated that half of companies that cut customer service staff due to AI would rehire by 2027. The pattern that’s emerging: cut fast, learn the hard parts, add humans back for the work AI can’t handle reliably.
Meanwhile, KPMG’s Q1 2026 Global AI Pulse report found that while 22% of U.S. businesses are exploring agents and 14% are actively deploying them, only 9% are orchestrating multiple agents across workflows the way ClickUp claims to. What ClickUp is doing is rare. Whether it’s visionary or premature is still an open question.
The contrarian case against the 100x org
The skeptic’s version of this story goes like this: a company that hasn’t raised money since its 2021 peak-froth Series C lays off 22% of staff, calls it innovation, and dangles $1M salary bands that almost nobody will actually reach.
Critics on LinkedIn and tech forums argued that the restructuring risks weakening core business functions like customer understanding and user research. If your AI agent fleet handles support tickets faster but never surfaces the subtle patterns a human support team would catch, you’ve traded speed for signal. That’s the Klarna lesson.
There’s also the 100x expectation itself. If companies expect fewer workers to produce dramatically more output with AI tools, those workers may face constant performance monitoring and unrealistic pressure. The math only works if the agents actually deliver. A 2026 TechCrunch-covered study by AI training firm Mercor evaluated agents from top-tier models on 480 workplace tasks. Every agent failed to complete most of its duties.
Gartner predicted autonomous business will be a net-positive job creator by 2028 to 2029, driven by new work that AI can’t absorb. The short version: the 100x org might be directionally right but too early. If your agent fleet works well enough to justify cutting 22% of headcount today, you’re ahead. If it doesn’t, you’ve gutted institutional knowledge that takes years to rebuild.
What founders should take from this
You don’t have to build a 100x org to learn from one. Here’s what the ClickUp experiment actually shows, whether it succeeds or not.
The AI-first hiring filter is already standard. Shopify proved it, ClickUp made it structural. If you’re hiring in 2026 and not asking “can an agent do this?” for every role requisition, you’re already behind the market. That doesn’t mean cutting people. It means being honest about which work is agent-ready and which isn’t.
Agent governance is a real function now. ClickUp’s agent org chart, cost tracking, and escalation protocols aren’t overhead. They’re the only reason 3,000 agents haven’t created chaos. If you’re deploying more than a handful of AI agents, you need someone responsible for each one. That’s what “system manager” means in practice.
The Klarna reversal is required reading. Moving fast with agents is good. Moving so fast you have to rehire humans 18 months later because CSAT tanked on complex tickets is expensive. The line between AI transformation and AI washing is whether the output actually improves.
Per-employee output is the metric that matters. Not headcount. Not agent count. Revenue per employee, customer outcomes per team member, product velocity per engineer. That’s what validates the model. ClickUp’s numbers for the next 12 months will either prove Evans right or join the Klarna file.
As we’ve tracked across SoftBank’s $75 billion France AI investment and Perplexity’s IP lawsuit with CNN, the AI landscape is moving fast in every direction. The 100x org is the boldest public statement any SaaS CEO has made about what AI means for company structure. It might be the future. It might be an overcorrection. But the underlying question, what happens when your agents can actually do the work, isn’t going away. Every founder building a team in 2026 is answering it whether they frame it that way or not.
