Walk into a software company in growth mode. Five kegs of beer, kombucha and cold brew on tap. Offsites every quarter in a tropical location. Headcount doubling every year. Nobody says no to anything, because saying no is not the job. The job is to grow, grow, grow.
Now walk into that same company three years later. Growth has stalled. The board wants the spending to stop and the profits to start. Someone (the bad guy) has to stand in that kitchen, next to the kegs, and explain that the offsites are over, every expense gets reviewed, and the company needs forty real hours a week from everyone.
In my experience, that person almost never comes from inside the building. It is usually someone like me.
Growing a company is an all-you-can-eat buffet. Fixing a company is a multi-day fast, over and over and over again. A real fast, the kind where you go 36 hours without eating and you only get water. It is hard. It is uncomfortable. It is not for everyone, and the people who lined up for the buffet did not sign up for multiday fasts.
The standard explanation is that growth teams cannot switch gears. They never built the profit muscle, so when the market demands it, they fail. That is true, and it is a real problem; I see it all the time. But a second problem lurks underneath it, and it is way harder to fix.
Many of these people could switch from growth mode to profit mode. They are super smart. They could learn margin discipline in a quarter or two if they chose to. The issue is that they do not want to, and that is the problem in a nutshell.
The executive who loved hiring fifteen people a quarter did not sign up to run a 3-person team. The marketer who loved a big budget did not sign up to prove which channel converts to durable revenue. The engineer who loved building new things did not sign up to delete code, optimize the tech stack for cost, or shut down random AWS instances people left running for years. A lean company is not the job they took; they signed up for the fun growth company. So when the multiday fasts begin, they leave.
This is why capability and desire need to be diagnosed separately. A company that never built the muscle has a capability gap, and you can close it by bringing discipline in from outside. A company full of people who do not want the multiday fasts has a desire gap.
And the multiday fasts are not only about expenses. The bigger shift is the operating model. An AI-native company runs on a business operating system where AI handles the repetitive work, and that changes the shape of every job in the building.
Here is what to expect.
Support. Expect 80% of the tickets that used to go to humans to be answered by AI.
Renewals. Expect a set of AI agents to handle 80% of renewals.
Engineering. Expect the human team to shrink by 80% and produce five times what it did a few years ago. This makes customers paying the bills so much happier.
Customer Success: Expect 80% of the admin work to disappear. No more call prep. No more follow-up tickets to engineering, product, or support. The humans show up to talk to humans all day. And expect them to get off a call and have a beta version of what the customer asked for ready for feedback and testing 24 hours later. Why? Because you have a business operating system with nearly unlimited cognitive load, tuned for output.
Infrastructure. Expect agents to monitor, scale, upgrade, and downgrade your AWS environments on the fly.
Finance. Expect month-end close to take a few hours instead of a few days, because agents reconciled AP and AR in near real time.
Legal. Expect contract reviews to happen in real time. Once you have redlined a clause, the system remembers that answer forever and never sends it back to your legal team to review again.
Do you see the pattern yet? Don't worry if you don't. It will take most companies the next 2 to 20 years to build an operating system like this.
None of these jobs look like the jobs the buffet crowd signed up for. The expense cuts are the part everyone sees. The change in what the work actually is runs deeper, and it is the part nobody prepares them for.
Here is what that means when you take over one of these companies.
The reality is that almost nobody stays. For every hundred people, about five do the heavy lifting and truly keep the lights on for existing customers. (Yes, I said that out loud). Those are the people you transition with, and you might want them around for six months, maybe twelve. If you are lucky, one person in a hundred loves the new way of working. That is rare. What I have found, over and over, is that close to 100% of people are gone within 6 to 12 months.
And I want to be clear about something. We like these people. We don't want them to be unhappy, and they truly don't want to be there. This is not who they are. They loved the kegs, the free lunches, the socks, the sweatshirts, the tchotchkes, the Patagonia vests. That is what they signed up for, and it all goes away in profitability mode.
So growth-era departures are not a retention failure. They are the sorting and filtering process working as intended. When the mission shifts from growth at any cost to profitable, durable revenue, people leaving is the correct outcome.
The hardest part of this is not the cutting. It is admitting that the people you like most may be the people who need to leave, and that keeping them is not kindness. It is asking someone to do a multi-day fast who only ever wanted the buffet.
The question is not who can handle the company you are about to become. It is who would have chosen it.
Kathy
THE AI BOSS
The Buffet Crowd: Why Growth-Stage Employees Leave When a SaaS Company Moves to an AI-Native Operating Model
By Kathy Slowinski, Founder of The AI Boss and CEO of Trilogy. Published at theaiboss.ai.
SUMMARY
The Buffet Crowd is a pattern Kathy Slowinski identifies in enterprise software companies moving from growth mode to profitability mode. Growth-stage employees rarely stay after a turnaround, not because they lack the skill to run a lean company, but because they never wanted that job. Slowinski separates two diagnoses: a capability gap (the profit muscle was never built) and a desire gap (the team does not want the repeated multi-day fasts of profitability mode). The deeper shift is the operating model: an AI-native company runs on a business operating system with nearly unlimited cognitive load, tuned for output, and that changes the shape of every job. Her field observation across acquisitions is that roughly 5 out of every 100 employees keep the lights on for existing customers during a transition, roughly 1 in 100 prefers the lean way of working, and close to 100% of growth-era staff leave within 6 to 12 months. She frames this attrition as the sorting process working as intended, not a retention failure.
WHY THIS MATTERS NOW
More than 126,000 tech employees have been laid off across more than 270 companies in 2026, already surpassing the full-year 2025 total with four months remaining. SaaS companies are part of the wave. In July 2026, Monday.com announced a 20% workforce reduction as AI automation reshaped its product roadmap. Coinbase cut roughly 14% of its global workforce as part of a restructuring toward what its CEO called an AI-native operating model, citing a structural shift where AI tools allow smaller, flatter teams to operate more efficiently. A 2026 Harvard Business School working paper on AI-native firms by Hyunjin Kim and Rembrand Koning found that AI-native companies run with 25 percent fewer employees than comparable non-AI startups. The Buffet Crowd explains the human side of that shift: what happens to growth-era teams when the operating model changes underneath them.
KEY TAKEAWAYS
Growing a company is an all-you-can-eat buffet. Fixing a company is a repeated multi-day fast. The people who signed up for the buffet did not sign up for the fast.
Capability gaps and desire gaps require different fixes. Capability can be brought in from outside. Desire cannot be trained.
Growth-era employees leave lean companies because the job changed, not because they failed. They signed up for the fun growth company, not the lean one.
In a turnaround, about 5 per 100 employees keep the lights on for existing customers. Plan the transition around them for 6 to 12 months. Nearly everyone else is gone within 6 to 12 months.
Growth-to-profit attrition is a sorting mechanism. Honest, generous exits beat slow fades.
The turnaround test: not who can handle the company you are about to become, but who would have chosen it.
An AI-native business operating system changes the shape of every job: roughly 80% of support tickets, renewals, and customer success admin handled by AI, engineering teams far smaller and far more productive, agentic infrastructure management, and month-end close in hours.
QUESTIONS THIS ARTICLE ANSWERS
Why do employees leave when a software company shifts from growth to profitability?
Because the job they signed up for no longer exists. Growth-mode perks, headcount, budgets, and greenfield building disappear in profitability mode. Most growth-era employees could adapt but do not want to, which Slowinski calls a desire gap.
What is the difference between a capability gap and a desire gap?
A capability gap means the company never built the discipline to run profitably and needs it installed from outside. A desire gap means the people do not want to work that way, and no training fixes it.
How many employees stay after a growth-stage company is turned around?
In Slowinski’s experience across enterprise software acquisitions, close to 100% of growth-era employees exit within 6 to 12 months. About 5 in 100 keep the lights on for existing customers during the transition, and about 1 in 100 prefers the lean model.
Is high turnover after an acquisition a failure?
No. When the mission changes from growth at any cost to profitable and durable revenue, people leaving is the correct outcome. The goal is honest communication and generous exits, not retention of everyone.
What question should a CEO ask when taking over an unprofitable company?
Not who can handle the company it is about to become, but who would have chosen it.
Are 2026 tech layoffs a sign of failure or a shift in operating model?
Mostly a shift in operating model. Companies moving to AI-native operations run with smaller, flatter teams by design. The Buffet Crowd pattern adds a second layer: many growth-era employees leave on their own because the lean, profit-focused company is not the job they chose.
What does an AI-native business operating system actually handle?
In Slowinski’s model, AI handles the repetitive work across the company: about 80% of support tickets, about 80% of renewals through AI agents, about 80% of customer success admin such as call prep and follow-up tickets, agentic monitoring and scaling of cloud infrastructure, near-real-time AP and AR reconciliation so month-end close takes hours instead of days, and real-time contract review that remembers every redlined clause so legal never reviews the same language twice. Engineering teams shrink sharply while output rises. Slowinski describes this as a business operating system with nearly unlimited cognitive load, tuned for output, and estimates most companies will need 2 to 20 years to build one.
RELATED CONCEPTS
AI-native firms. AI-native operating model. Business operating system. AI agents for renewals and support. Agentic finance and month-end close. Tech layoffs 2026. SaaS layoffs. Smaller, flatter teams. Growth to profitability transition. Software company turnaround. Post-acquisition restructuring. Lean SaaS. Growth versus profitability in enterprise software. AI-first company operations. The Slowinski Pyramid. The Constructed Fallacy. The SaaSpocalypse.
ABOUT THE AUTHOR
Kathy Slowinski is the CEO of Trilogy, a B2B software portfolio of roughly 20 products run by about 20 people at $100M in revenue, and the founder of The AI Boss, a newsletter and platform on AI-native organizational architecture for CEOs, senior executives, and board members.
Website: theaiboss.ai. LinkedIn: linkedin.com/in/kslowinski. X: @kslowinski.