In my last post, The Four Stages of AI in Business, I laid out the framework I think every white-collar industry is walking through right now.
Stage 1: no AI.
Stage 2: humans doing the work, assisted by AI.
Stage 3: AI doing the work, reviewed by humans.
Stage 4: AI doing the work, no humans.
Since publishing that post, almost every business owner I've talked to has asked me the same question: "How do I move my company from Stage 2 to Stage 3?"
My honest answer is that they probably can't.
Not because they're not smart enough. Not because they're not willing to work hard. But because the jump from Stage 2 (human-first, AI-assisted) to Stage 3 (AI-first, human-reviewed) isn't a step. It's a cliff. And the businesses that try to leap it from the wrong side are almost guaranteed to lose to brand-new AI-first companies that started life on the other side.
This isn't a new pattern. When I wrote The Disruption Formula, I studied more than 25 case studies of companies that disrupted their industries. The overwhelming majority were new entrants. Almost none were existing players who successfully reinvented themselves. Blockbuster didn't become Netflix. Kodak didn't lead the digital camera era. Sears didn't become Amazon. The incumbents almost never make the leap, because the very things that made them successful in the old model become the anchors that drag them under in the new one.
AI is about to run the same playbook on the entire white-collar economy.
Let me explain why, using a tax firm as the example. The numbers are made up but the math reflects what's actually happening across industries right now, from legal to bookkeeping to healthcare to recruiting.
Problem 1: Your Workforce Gets Cut by 90%
A traditional, Stage 2 tax firm files 100 returns a year with 10 employees. Each preparer handles about 10 returns. They use AI tools to go faster, but they're still doing the work themselves.
A Stage 3 AI-first tax firm files the same 100 returns with one employee. The AI agent prepares all 100 returns. The one human reviews and signs off.
Read that again. Same revenue, same client base, same volume of work. Nine fewer people.
That's not a "transition." That's nine families who depended on a paycheck no longer having one. It's nine people the owner hired, trained, mentored, and may have known personally for years. Most owners physically cannot bring themselves to do this. So they don't. They keep all ten employees, keep doing it the old way, and hope the future doesn't arrive.
A new AI-first competitor doesn't have this problem. They never hired the nine in the first place.
This is the "seat compression" effect now playing out across the entire SaaS and professional services world. One AI agent replaces what used to be ten human seats. It's already happened in software engineering. It's happening now in tax, legal, bookkeeping, customer support, and recruiting. It's coming for every white-collar industry next.
Problem 2: Your Pricing Power Collapses
A traditional CPA might charge $1,000 to prepare and file a return. That price made sense when ten hours of human time went into it.
In an AI-first model, ten minutes of human review time goes into it. The rest is software. The cost to produce that return has dropped by 90% or more.
In theory, the AI-first firm could keep charging $1,000 per return and just pocket the margin. And for a little while, some of them will. But the moment a competitor shows up charging $499, or as Deduction.com is already doing, $499 for unlimited tax help all year, the market price collapses to the new floor.
Jeff Bezos has a line I think about a lot: "Your margin is my opportunity." The AI-first firms that try to keep traditional pricing are signaling exactly where the next competitor should attack. The new entrants will gladly take 30% margins on $499 returns and put the $1,000 incumbents out of business.
So now the owner is staring at: 90% fewer employees and roughly half the revenue. That's the actual shape of the AI transformation most business owners aren't ready for.
A new AI-first competitor doesn't have this problem either. They built their cost structure around $499 from day one. To them, that's not a discount. That's the business model.
Problem 3: Everything Else Has to Be Rebuilt
The office space was sized for 10 employees. Now it's sized for 1. The lease isn't.
The management team existed to manage the 10 employees. With one employee, there's nothing to manage.
The HR systems, the training programs, the career ladder, the bonus structure, the org chart, the onboarding playbook, all of it was designed for a firm with people in it. None of it applies to a firm where the AI does the work.
The financial projections, the bank loans, the credit lines, the partner draws, all sized for $1,000 returns and 10 employees of labor cost. None of it survives contact with $499 returns and one employee.
Even the sales process changes. You used to sell "our expert CPAs will prepare your taxes." Now you have to sell "our AI prepares your taxes and a CPA reviews them," which is a completely different pitch to a completely different customer.
A new AI-first competitor doesn't have any of this baggage. They sign a small lease, hire one reviewer, build the AI workflow, and start filing returns. While the incumbent is unwinding a fifteen-year-old org chart, the new entrant is shipping.
So What Should You Actually Do?
If you own a Stage 2 business, this is the conversation I find myself having again and again. There are really only three paths:
Option 1: Do nothing.
Stay in Stage 2 as long as you can. Ride it out. Take cash off the table while it lasts. Eventually an AI-first competitor takes your customers and you fade into oblivion. Some owners genuinely choose this path, and depending on your age and your goals, it isn't always a bad call. But it's a managed decline, not a strategy.
Option 2: Try to transition.
Take your existing company and rebuild it into an AI-first firm. This is the path most owners say they want to take, and it's the one I'm most skeptical of. You'd be unwinding ten years of decisions while an AI-first startup, with no baggage, races past you. You'd spend 99% of your energy fighting inertia inside your own company while they spend 100% of theirs on product, customers, and growth. Tooth and nail, every day, for years.
Option 3: Sell, then start fresh.
Sell your Stage 2 business while it still looks like a healthy traditional firm to a buyer who hasn't internalized what's coming. Take that capital, plus the relationships and industry knowledge you've built, and start a new AI-first company from scratch. No nine employees to fire. No $1,000 pricing to defend. No fifteen-year-old org chart. Just you, the AI, and a clean slate to build the thing the right way from day one.
This is the debate I'm having in my consulting work with business owners right now. Most of them desperately want Option 2 to be the answer because it lets them keep what they've built. But the more I think about it, the more I believe Option 3 is the only path that actually plays to a founder's strengths instead of fighting against them.
The work you'd have to do to transition an existing company into an AI-first model is roughly the same work you'd have to do to build a new one. Except in the new one, every hour goes into the future. In the old one, most of your hours go into dismantling the past.