The hardest decision Roy Mann has made since taking monday.com public isn't about headcount, pricing, or product roadmap.
It's about a screen. Specifically: the first screen a new customer sees when they sign up.
For years, that screen showed a board - a visual representation of your workflow. Intuitive, familiar, immediately useful. It's a big part of why monday.com has 250,000 paying customers across 200 industries.
They've changed it. New users now onboard onto an agent.
That single decision cascades through the entire funnel: acquisition, activation, retention, revenue. It's not a feature change. It's a statement about what the product fundamentally is. And Roy told me on the AI Revolution Show last month that it was the hardest call he's had to make, precisely because everything was still working.
"The hardest thing is letting go of something, especially when it's growing. That's a hard decision to make."
That sentence deserves to sit with every founder reading this. Not when you're losing. When you're winning.
THE REBUILD REPORT
monday.com in numbers, as of early 2026:
THE BEFORE STATE
Monday's original vision was to manage all work. It was an intentionally wide problem: not project management, not task tracking, but how people work together. That breadth is what made beer breweries and clinical trial researchers and airplane manufacturers all find a home on the platform.
The business model matched: per-seat SaaS. You grew by adding users. The more seats, the more revenue. Clean, predictable, scalable.
Then two things happened simultaneously. AI agents started doing work that humans used to do - shrinking the headcount that per-seat models depend on. And customers started expecting software to do the work, not just organise it.
Roy's read on the market was stark:
"SaaS is being valued to basically nothing because investors don't know who's going to make it or not. It's a massive change, so they pull out of everything at once."
He's not panicking. He's describing a structural re-rating. The market doesn't know which SaaS companies will survive the transition, so it's discounting all of them until the picture clears.
THE DECISION
Monday changed their vision statement. It now reads: not just manage work, but do the work. Three words added. Everything underneath them changed. The practical expression of that change came in three specific moves:
1.New users onboard onto an agent, not a board.
The funnel is rebuilt around AI-first value, not workflow visualisation.
2.Agents can now buy seats on the platform.
Monday opened the product to AI agents as customers, not just tools. Monday vibe, monday agents, monday sidekick all launched within 12 months.
3.‘Agent Talent’ launched.
A marketplace where any agent can sign up, offer services and complete for work inside customer organisations. Roy described it as a ‘marketplace for jobs, for agents’.
WHAT MOVED
monday vibe hit $1M ARR in 2.5 months - the fastest product in monday's history to reach that milestone. That's the early signal that the AI-native layer has genuine customer pull, not just internal belief.
But Roy was unusually candid about what hasn't moved yet:
"I can tell you how much money I spend on tokens and what do I get for it - it's not a good ratio. But I see what it can do."
That honesty matters. A $1.5bn-cash public company CEO saying the unit economics of AI aren't working yet, but building anyway, is telling you something about conviction versus certainty. He's not building because the numbers are there. He's building because he believes they will be, and that waiting for them guarantees you miss the window.
The agent seats are signing up, but Roy's own words: not significant yet. The market is pre-chasm. The bet is on who builds the infrastructure before the chasm arrives.
THE LESSON
Monday's rebuild is not a story about a company in trouble making a desperate pivot. It's a story about a company that's growing, profitable, and cash-rich - choosing to make hard changes anyway.
Roy went back to writing code. Not because he had to. Because he believed you cannot drive change in technology you don't understand deeply. That's a personal operating decision that most CEOs at his scale don't make.
The question it leaves for every founder reading this isn't 'should I rebuild?' It's Roy's sharper version:
"A lot of SaaS companies are solving a core problem, and maybe that problem is not changing. Maybe it's the same problem people will still have - but the way it's going to get solved is going to change fundamentally."
The problem your customers have probably isn't going away. The mechanism you use to solve it almost certainly is.
ONE QUESTION:
Monday changed the first screen new users see. What's the equivalent decision in your product: the one that would change your entire funnel if you made it?
You don't have to have made it yet. Hit reply: the founders who can name that decision clearly are usually the ones closest to making it.
Alex
Founder, Shift AI
P.S. Roy's full conversation is on the AI Revolution Show - the podcast is live now wherever you listen. He's one of the few public company CEOs willing to say out loud that the economics aren't working yet while making the bet anyway. That combination of honesty and conviction is rare.
The Shift · Published every Wednesday · shift.ai
You're receiving this because you subscribed. Unsubscribe any time.