75% of current CPAs are nearing retirement. Accounting degrees dropped 7.4% last year — the steepest decline in 30 years. Only 1.4% of college students choose accounting, down from 4%. The profession is projecting a 140,000-accountant shortfall by 2027 with 136,400 annual job openings through 2034. The supply of practitioners is shrinking. The supply of buyers for their practices is shrinking with it.
Median firm valuation sits at 1.0x revenue. But that median hides a widening split. Private equity has poured $50 billion into accounting in six years, with deal flow doubling. Capstone Partners reports top firms commanding up to 15x EBITDA. Meanwhile, compliance-heavy practices with high client concentration face valuation haircuts of 20–40%. The ceiling is rising. The floor is dropping. This is a K-shaped valuation market — and most practice owners approaching retirement don't know which side they're on.
The K-shaped split
I was talking to a new associate recently who helps accounting firms sell their practices. The Sequoia memo on AI disrupting professional services had come out the week before, and I told him what I've been thinking for months: the valuation model for accounting firms is about to get fundamentally altered.
The data backs this up. PE firms aren't buying client lists. They're buying systems — encoded knowledge, advisory capacity, AI-integrated delivery, recurring revenue models. Firms with 80%+ recurring revenue command multiples 0.2–0.4x higher. Advisory mix commands premiums. AI integration is now influencing deal pricing more than traditional book-of-business metrics. Firms using AI report 37% higher revenue per employee. 63% of buyers prefer non-hourly billing models.
On the other side of the K, a new practitioner with cloud tools and AI can start a practice for roughly $20,000 in annual technology investment and replicate what previously required five staff. Why pay 1.0x revenue for a firm with legacy systems and institutional knowledge locked in the departing owner's head when you can build the same client capacity from scratch in 18 months?
You've seen this movie before — but the ending is different
We saw a version of this during the cloud adoption wave. Younger practitioners bought practices from retiring owners, planning to capture efficiency gains by migrating to Xero or QBO. Some succeeded. Many found the client base was as resistant to change as the outgoing owner — built on acquiring new clients rather than increasing the value of existing ones. The economics still worked because cloud efficiency was yours to keep. You moved to the cloud, your delivery cost dropped, your fees stayed the same. The margin improvement justified the acquisition.
The AI wave is different. Clients expect fees to drop because they can see the efficiency. Bloomberg reported PwC clients demanding AI discounts — arguing that if the work is faster, the savings should flow to them. This isn't optional efficiency you capture as profit. This is forced efficiency where clients expect lower fees.
That changes the valuation math entirely. A buyer paying 1.0x today's revenue is paying 1.0x revenue that's about to compress. The effective multiple is higher than it appears. Valuation models built on current monthly recurring revenue and existing profit margins are pricing a future that won't exist.
The knowledge that walks out the door is the knowledge you couldn't sell
Here's the question a buyer is really asking: what am I purchasing that survives the founder's exit?
Post-acquisition, 50% of key employees leave within the first year. 75% leave within three years. If a practice's operational knowledge lives in the owner's head and the senior bookkeeper's muscle memory, the buyer isn't purchasing a firm. They're purchasing a client list with no delivery system attached. When those people walk, the buyer is left with names and no way to serve them.
A practice whose knowledge lives in structured systems — SOPs, client profiles, workflow automation, encoded decision rules — has something transferable. The knowledge survives the transition. The delivery system works regardless of who's running it. That's what commands a premium.
This is context engineering applied to exit strategy. The difference between a sellable firm and an unsellable one isn't revenue, client count, or even advisory mix. It's whether the firm's intelligence is encoded in systems or locked in people. One is an asset. The other is a liability disguised as goodwill.
Your exit plan is a context engineering problem
The valuation improvement timeline is 18–36 months. If you're 3 years from exit and your firm's knowledge lives in your head, you're running out of runway.
The practices that encode their knowledge, build advisory capacity, and integrate AI into a transferable delivery system will sell at a premium in this market. The practices that don't will discover that a client list without encoded knowledge is a depreciating asset — and the buyer who might have saved it can now build their own from scratch for less.
If you're building toward an exit in the next 5–10 years, the question isn't what your firm is worth today. It's which direction the value is moving — and whether you have time to change the trajectory. Let's look at what a buyer would actually be purchasing. Visit theaiaccountant.ai/consultation to book a conversation.