The Margin Illusion

AI is not destroying margins. It is destroying the illusion that operational efficiency was a competitive advantage.

AI is turning PE's best margin play into everyone's margin play.

Private equity's dominant value creation model has run on the same engine for three decades. Buy a business with operational inefficiency. Install professional management. Reduce costs. Optimise processes. Improve margins. Exit at a multiple of improved EBITDA. Repeat.

The returns have been extraordinary. And the model worked because it rested on three assumptions that were reliably true: operational execution was expensive, labour-intensive, and difficult to replicate at speed. A PE firm that could professionalise operations faster than the market could copy them had a genuine window of alpha.

AI is about to make all three assumptions false simultaneously.

The data is already visible

Britain's professional services sector accounts for 81% of GDP. 83% of British workers are in services, the highest proportion of any large rich economy. Morgan Stanley analysis suggests the UK has already suffered the greatest net AI-driven job losses among comparable economies. Goldman Sachs research confirms that employers are responding by hiring less and investing more in capital assets, substituting human labour for automated capability.

This is not a forecast. It is a current account.

At Charles Russell Speechlys, a trainee solicitor takes two days to review and summarise a hundred-page legal document. AI produces equivalent output in ten minutes. PwC has already cut hundreds of graduate positions. Finance, insurance, and information and communication experienced the sharpest employment declines in the year to September 2025, contributing to UK unemployment reaching a five-year high of 5.2%.

These are not isolated disruptions in discrete sectors. They describe a structural shift in the cost of professional execution. Every function that PE firms currently optimise in portfolio companies is experiencing the same dynamic: AI compresses the cost of execution toward zero.

Research on AI's theoretical capability across occupational categories confirms the scale of this shift. Across management, business and finance, legal, and office administration, AI can theoretically perform the vast majority of job tasks. Observed adoption remains a fraction of that capability. The gap between what AI can do and what organisations are currently using it for is enormous, and it is closing. When it does, every operational improvement PE has made in portfolio companies becomes replicable by any buyer with the same tools and a quarter of focused implementation.

When the cost of doing things well falls to near-zero, the value of doing things well falls with it. The question is no longer "can you execute efficiently?" The question is "why should anyone pay you a premium for execution that everyone can now do?"

The illusion

Here is how it plays out in practice. A PE firm acquires a professional services business. It installs AI across operations. Headcount drops 30%. EBITDA improves dramatically. The operating partner presents this to the investment committee as value creation.

It is not value creation. It is value revelation. The margin was always there, hidden behind labour costs that AI has now made optional.

And here is the problem: every competitor in the sector can do the same thing, on the same timeline, with the same tools. The models are public. The infrastructure is commoditised. A five-person firm has access to the same AI capability as a global network. Scale used to be the moat. Now it is a tax.

This is the alpha-to-beta shift. When an operational improvement is available to every market participant simultaneously, it ceases to be a source of competitive advantage. It becomes table stakes. AI-driven cost reduction is not proprietary. It is not defensible. It is not a reason for a buyer to pay a premium at exit.

Which creates the compression trap. Portfolio companies that compete primarily on operational efficiency will find themselves in a race to the bottom as AI enables every competitor to achieve the same cost structure. Margins improve briefly, then compress as the market equilibrates. The PE firm's exit multiple reflects not the improvement but the market's recognition that the improvement is replicable.

Buyers at exit are not stupid. If the margin improvement came from AI-driven cost reduction that any acquirer could replicate with a licence and 90 days, it commands no premium. The multiple reflects defensibility. AI efficiency is not defensible.

The question at exit becomes: what does this business have that a competitor cannot replicate? For almost every PE-backed business, the answer is brand. Or it is nothing.

What sits above the automation line

Think of it as a line drawn across the capabilities of any business.

Below the line sits everything AI can do as well as or better than humans: document review, data analysis, process optimisation, content production, customer service triage, financial modelling, code generation, scenario planning, administrative coordination. This is where the majority of PE operational improvement has historically focused. And this is what is being commoditised.

Above the line sits everything AI cannot do: judgment, taste, relationships, cultural understanding, strategic positioning, narrative, identity, trust, meaning. These capabilities are far harder to automate because they depend on context, relationships, accumulated history, and the kind of cultural meaning that machines cannot replicate.

For 30 years, PE value creation focused on optimising everything below the line. This was true when below-the-line excellence was scarce and hard-won. It is no longer true when AI makes below-the-line excellence universal.

The inversion is now required. Value creation must start above the automation line and work downward. Build the brand, the positioning, the narrative, the meaning. Then use AI to execute against it with ruthless efficiency. The strategy precedes the automation, not the reverse.

In law, AI commoditises document review and conveyancing. The firms that survive at premium pricing are those where the brand represents judgment, not process. In consulting, AI commoditises data analysis and scenario modelling. PwC's own head of consulting says it plainly: the premium on judgment, creativity, and insight is rising. In consumer businesses, AI commoditises production, logistics, and customer service. The brands that justify premium pricing are those where the meaning of the product exceeds its functional value.

The automation line is the new dividing line in PE value creation. Everything below it is converging. Everything above it is diverging.

The blindspot

Why have PE firms historically undervalued brand? Because it is intangible. Because it is difficult to measure in a way that survives due diligence. Because operating partners are trained in finance, operations, and commercial strategy. Very few have brand expertise. The result: brand is treated as a marketing cost, not a value creation lever.

This was an acceptable blindspot when operational efficiency was the primary source of alpha. It is no longer acceptable when operational efficiency is being commoditised.

But the blindspot runs deeper than organisational habit. The real problem is that PE firms measure brand as a cost rather than as what it actually is: a risk-reduction instrument.

Ian Whittaker's IPA research, based on interviews with more than 200 analysts and investors, found that 75% believe marketing should be treated as investment, capitalised either in full or in part. His central argument is that the most powerful case for marketing is not that it drives growth. It is that it reduces risk. A business with stable demand, loyal customers, and pricing power has higher-quality earnings. Higher-quality earnings deserve a higher valuation multiple. That is not marketing theory. That is how capital markets work.

The implication for PE is direct. Two portfolio companies with identical EBITDA but different brand strength have different earnings quality. The one with the brand moat has more repeatable revenue, more defensible margins, and lower demand volatility. At exit, that difference shows up in the multiple. Not because the buyer "values the brand" in some abstract sense, but because the buyer's DCF model applies a lower discount rate to more predictable cashflows.

Most PE operating partners cannot articulate the brand strategy of their portfolio companies. They can articulate the commercial strategy, the operational strategy, the financial strategy. They cannot explain why a customer chooses their portfolio company over a competitor in a single sentence. That gap is where value is leaking. AI is about to widen it into a chasm.

What to do about it

Brand audit at acquisition, not after. Due diligence currently evaluates financial performance, operational efficiency, market position, and management quality. It should also evaluate brand equity: pricing power, customer loyalty drivers, narrative clarity, competitive positioning. A business with strong brand equity and weak operations is more valuable in an AI world than one with weak brand equity and strong operations. AI fixes operations. Nothing fixes a meaningless brand except strategy and time.

Brand on the value creation plan, not the marketing plan. Brand development should sit alongside operational improvement, commercial strategy, and management upgrades as a named value creation lever with its own KPIs, timeline, and accountability. It is not a cost centre. It is a value driver.

Above-the-line capability as a hiring criterion. Operating partners and portfolio company leadership need people who can do the above-the-line work: strategic positioning, narrative development, cultural insight, meaning-making. These are not marketing hires. They are strategic hires.

Measure brand as a risk-reduction instrument, not a growth driver. Track earnings volatility, demand repeatability, pricing power resilience, and customer concentration risk as brand KPIs. These connect brand investment to the capital allocation framework CFOs actually use. As Whittaker argues, the CFO who measures brand as a cost will always cut it under pressure. The CFO who measures brand as a risk-reduction instrument will protect it under pressure, because cutting it increases earnings volatility, which increases the risk premium, which decreases enterprise value. That is not a marketing argument. That is a finance argument.

Use AI below the line to fund investment above it. The efficiency gains from AI-driven operational improvement should not flow entirely to EBITDA. A portion should be reinvested in above-the-line capabilities: brand, positioning, narrative, customer meaning-making. The firms that extract AI savings and reinvest in brand will compound value. The firms that extract and distribute will find the savings competed away within 18 months.

The convergence thesis

We are entering a period of unprecedented convergence in execution capability. AI gives every business access to the same tools, the same efficiency, the same cost structure. In that world, the only remaining source of divergence is meaning: why your business exists, what it stands for, and why anyone should care.

Private equity has spent 30 years getting extraordinarily good at making businesses run better. The next 30 years belong to the firms that get extraordinarily good at making businesses mean something.

When execution costs converge, meaning diverges. Brand is how you create meaning. Meaning is how you justify margin. Margin is how you create value. The firms that understand this first will not just outperform. They will define the category.