An owner I sat with this spring had just come out of his first conversation with a corporate finance adviser. The adviser had been polite about it. The firm was profitable, the team was capable, the client list was solid. The number on the table at the end of the meeting was still lower than he had expected, and the reason was the one sentence everyone in the room had been circling around. “The firm is too dependent on you.”
He told me afterwards he had nodded through the conversation, gone home, and spent the weekend doing the arithmetic. The gap between what he had been working towards for twenty years and what was actually on offer was real, and the discount had a name even if he had not heard it before.
Three to five years out from a planned exit is the window where this conversation matters. Inside twelve months it is largely too late, and beyond five years it is too easy to defer. The question is what to do with the time you have, and how much of the picture AI now changes.
What is the founder-dependency valuation discount?
The founder-dependency valuation discount is the gap between what a buyer pays for a founder-dependent services firm and what they pay for an otherwise identical founder-independent peer. In the UK lower-middle-market, that gap sits at 30 to 50 per cent. In multiple terms, founder-dependent firms struggle to clear 3 to 4 times EBITDA while founder-independent peers in the same sector at the same earnings command 7 to 8.
The number is documented across several independent sources. Strategic Exit Advisors put the range at 30 to 50 per cent based on their transaction work. William Buck, working from the key-person-risk side of chartered accountancy practice, puts the pure key-person discount at 10 to 25 per cent, which suggests the full 30 to 50 per cent includes compounding factors beyond key-person risk alone. SME Business Valuation, looking specifically at UK firms in the £3m to £30m band, reports 20 to 40 per cent. The ranges differ but the direction is the same, and the discount is structural rather than negotiable in the room.
Why does it matter for your business in a way it did not before?
It matters because the discount decomposes into specific components, and AI shifts some of them. The aggregate 30 to 50 per cent breaks down into customer concentration, lack of documented systems, management depth, recurring revenue mix, sales systematisation, and financial reporting quality. Each is priced separately by a buyer’s diligence team. Some of those components live in process and can now be encoded into firm-owned assets. Others live in relationships and cannot.
The shift is most concrete on systems documentation, where Livmo and similar transaction analysis report a 20 to 40 per cent sale price uplift on firms with systematic SOP coverage. Three years ago, documenting a senior advisor’s drafting methodology was a manual exercise that competed for time with billable work. Today, the same advisor running their drafting through a prompt library and workflow assembly is producing transferable IP as a by-product of doing the work. The documentation happens because the system requires it, not because the owner found a fortnight to write it down.
The decision-framework side has shifted similarly. Pricing rubrics, scoping templates, and quality-control checklists that used to live in the founder’s head can now sit inside a tool the team uses every day. A buyer’s diligence team interviewing the management layer will hear the team articulate the methodology because the team uses it, not because they were briefed for the call. That is the kind of evidence that moves the management-depth component of the discount, and it was much harder to produce before.
Where will AI actually move the dial, and where will it not?
AI moves the dial on components of the discount that live in documented process, methodology, and decision frameworks. It does not move the dial on relationship capital, regulator trust, brand association with the founder, or judgement on subjective work. A simple test: could a senior team member, with access to your stack and documentation, deliver the same output to acceptable quality? If yes, AI is helping. If no, the dependency does not encode.
The components that move are concrete. Sales pipeline systematisation moves when lead qualification, proposal drafting, and scoping calculations sit in tooling the business development layer uses without founder intervention. Financial reporting quality moves when bookkeeping and management accounts are produced on a consistent cycle by a part-time CFO using documented processes rather than ad hoc by the owner. Process documentation moves when SOPs are a by-product of using AI-driven workflows. These are the components Anders CPA describes as scrutinised in a Quality of Earnings review, and they are also the components AI directly improves.
The components that do not move are equally concrete. A buyer’s reference calls to your top ten clients will tell them quickly which contacts stay for the firm and which stay for you specifically. A twelve-year relationship with a regulator or framework partner that runs through your personal credibility does not transfer to an AI workflow, no matter how well documented. The Built to Sell test (would the customer remain if the founder left) is the one that catches over-claiming, and a sale memo that suggests AI has replaced relationship capital reads as naive at the first diligence call.
When is the right time to act, and when is it too late?
The right time to start is when a three to five year exit becomes a serious consideration. The Exit Planning Institute’s Discover-Prepare-Decide framework treats 24 to 36 months as the value-acceleration window, and the components that move the multiple need most of it. Customer relationship transfer takes 18 to 24 months. Management depth takes a similar period. Recurring revenue conversion is a multi-quarter conversation, not a memo.
The final twelve months can address financial hygiene, documentation polish, customer reference preparation, and add-back normalisation for a Quality of Earnings review. What they cannot address is whether the firm operates without the founder, because demonstrating that takes time the buyer can see in the records. Axial’s analysis of 75 failed UK and US transactions in 2025 found that 25 per cent collapsed because of diligence findings outside formal QoE work, and the most common category was customer concentration and key-person risk surfacing too late to remediate. The pattern is consistent. Late starts get caught.
For a founder with three to five years on the clock, the productive sequence is to begin with the components AI now meaningfully assists with. Methodology documentation, workflow assembly, decision-framework codification, and pricing-rubric encoding can be in place within twelve months. That creates the platform for the longer-running work (management depth, customer relationship transfer, recurring revenue mix) across the remaining two to four years. The AI-encoded work makes the slower work easier, not the other way around.
Related concepts and the wider picture
The founder-dependency valuation discount sits inside a broader pattern about how owner-managed services firms create or destroy enterprise value. Customer concentration, recurring revenue mix, management depth, documented systems, and financial reporting quality each have their own treatment in the buyer’s diligence framework. Built to Sell anchors the customer-relationship dimension, the Exit Planning Institute frames the methodology, and Quality of Earnings analysis covers the financial side.
The AI shift is recent enough that the public literature on it is still catching up. What the operating experience shows is that encoding process, methodology, and decision frameworks into firm-owned tooling materially changes the management-depth and systems-documentation components, while leaving the relationship-capital and judgement components untouched. Both halves matter. Acting only on the encodable half leaves money on the table where direct human work is needed. Acting only on the human components misses the multiplier AI now provides on the process side.
If you want to talk through where your firm sits across these components and what the realistic three to five year picture looks like, Book a conversation.
