Reinventing Traditional Businesses Through Generative AI
Inflection Ventures partners with legacy service businesses - call centres, BPOs, managed services - to drive 10× value creation in less than 5 years, by shifting from labor-first to AI-first.
Margin Expansion vs Revenue Growth: Why Investors Are Rethinking the Formula
Margin Expansion vs Revenue Growth: Why Investors Are Rethinking the Formula
September 15, 2026
⏱ 9 min read
by Inflection Ventures
September 15, 2026︱⏱ 9 min read︱by Inflection Ventures
Shares
For years, the default answer to increasing enterprise value was straightforward: grow revenue. More customers, more contracts, more employees to deliver them, and more revenue on the top line. But that formula becomes less attractive when every additional dollar of revenue still requires another dollar of cost to deliver it.
That is the problem facing labour-first businesses today. Investors are demanding more EBITDA growth from the same five-year hold, while the traditional route to growth can require businesses to keep adding people, capacity and cost. A deal that once needed roughly 5% annual EBITDA growth to reach a 2.5x return over five years now needs 10% to 12% (Bain, Global Private Equity Report 2026).
The obvious response is to grow faster. But there is another route: make the existing revenue more profitable. Companies that have rebuilt their operating model already run about 20% higher revenue per employee than opportunistic AI adopters, without necessarily growing revenue any faster (McKinsey, Beyond Productivity: How AI Creates Value in Private Equity, 2026, 471 PE-backed companies with $1m to $250m revenue).
Two routes sit behind every higher valuation: grow revenue at the same cost structure, or expand margin at the same revenue. In a labour-first service business the two are not equally available, and most owners default to the one that costs more to attempt.
What you will find in this article:
1. Why revenue growth and margin expansion are not the same size of task in a labour-first business, and which one is structurally cheaper to pull
2. What each path actually costs to attempt, in headcount, capital and time
3. Which of the two paths you control directly, and which one depends on the market saying yes
4. Where operators get the choice backwards, and how the same framework reads from an advisory engagement versus an acquisition
For years, the default answer to increasing enterprise value was straightforward: grow revenue. More customers, more contracts, more employees to deliver them, and more revenue on the top line. But that formula becomes less attractive when every additional dollar of revenue still requires another dollar of cost to deliver it.
That is the problem facing labour-first businesses today. Investors are demanding more EBITDA growth from the same five-year hold, while the traditional route to growth can require businesses to keep adding people, capacity and cost. A deal that once needed roughly 5% annual EBITDA growth to reach a 2.5x return over five years now needs 10% to 12% (Bain, Global Private Equity Report 2026).
The obvious response is to grow faster. But there is another route: make the existing revenue more profitable. Companies that have rebuilt their operating model already run about 20% higher revenue per employee than opportunistic AI adopters, without necessarily growing revenue any faster (McKinsey, Beyond Productivity: How AI Creates Value in Private Equity, 2026, 471 PE-backed companies with $1m to $250m revenue).
Two routes sit behind every higher valuation: grow revenue at the same cost structure, or expand margin at the same revenue. In a labour-first service business the two are not equally available, and most owners default to the one that costs more to attempt.
What you will find in this article:
1. Why revenue growth and margin expansion are not the same size of task in a labour-first business, and which one is structurally cheaper to pull
2. What each path actually costs to attempt, in headcount, capital and time
3. Which of the two paths you control directly, and which one depends on the market saying yes
4. Where operators get the choice backwards, and how the same framework reads from an advisory engagement versus an acquisition
Shares
The Two Paths to a Higher Multiple
The Two Paths to a Higher Multiple
Revenue growth depends on the market saying yes. Margin expansion depends only on you.
Every higher valuation is arithmetic before it is a story. Enterprise value moves when EBITDA grows, when the multiple applied to it grows, or both. For an operations-heavy service business (a call centre, a BPO, a managed-services or field-engineering operation) there are really two levers available at the EBITDA line: sell more at the same cost structure, or spend less to deliver the same revenue. Everything else is a variant of one of those two.
The reason this choice matters more than it used to is that the bar has moved. A typical buyout in the 2010s needed roughly 5% annual EBITDA growth to reach a benchmark 2.5x return over a five-year hold. Bain's own description of the shift is blunt: deals now need "10% to 12%" annual EBITDA growth to hit the same target, because "generating attractive returns now requires significantly more operational improvement and revenue growth" (Bain, Global Private Equity Report 2026, published 23 February 2026).
Read as an owner rather than a buyer, that sentence contains a choice Bain does not resolve for you. "Operational improvement and revenue growth" are not one lever. They are two, and in a labour-first business they do not cost the same to pull.
This is where Inflection Ventures' approach starts to differ from a conventional “find an AI tool” exercise. The question is not: “Where can we use AI?”
Revenue growth depends on the market saying yes. Margin expansion depends only on you.
Every higher valuation is arithmetic before it is a story. Enterprise value moves when EBITDA grows, when the multiple applied to it grows, or both. For an operations-heavy service business (a call centre, a BPO, a managed-services or field-engineering operation) there are really two levers available at the EBITDA line: sell more at the same cost structure, or spend less to deliver the same revenue. Everything else is a variant of one of those two.
The reason this choice matters more than it used to is that the bar has moved. A typical buyout in the 2010s needed roughly 5% annual EBITDA growth to reach a benchmark 2.5x return over a five-year hold. Bain's own description of the shift is blunt: deals now need "10% to 12%" annual EBITDA growth to hit the same target, because "generating attractive returns now requires significantly more operational improvement and revenue growth" (Bain, Global Private Equity Report 2026)
Read as an owner rather than a buyer, that sentence contains a choice Bain does not resolve for you. "Operational improvement and revenue growth" are not one lever. They are two, and in a labour-first business they do not cost the same to pull.
This is where Inflection Ventures' approach starts to differ from a conventional “find an AI tool” exercise. The question is not: “Where can we use AI?”
It is: “Where does changing the way work gets done create the greatest measurable movement in the P&L?”
It is: “Where does changing the way work gets done create the greatest measurable movement in the P&L?”
What Each Path Costs to Attempt
Revenue growth in a labour-dependent business is rarely free of the cost base. Inflection Ventures' own diagnostic work across client operations describes the mechanism plainly: to double revenue, a labour-first operator typically has to roughly double staff, so cost grows in step with revenue and margin never improves with scale. That is a structural property of the operating model, not a management failing. It is what "labour-first" means.
Which makes revenue growth, in this kind of business, a capital- and time-intensive bet before it is a margin story. New revenue has to be sold, then delivered, which means hiring ahead of it, training into it, and carrying the ramp-up cost before the revenue lands and often for some time after. None of that is guaranteed to convert. It depends on demand existing, on winning it against competitors, and on execution holding through the ramp. The cost is committed before the outcome is known.
Margin expansion starts from a different position: the revenue is already there. The task is to redesign how the existing work gets done so the same output requires less cost, which is a bounded, scopeable piece of work rather than an open-ended one. It is also, in Inflection Ventures' own delivery data, a lever with a known shape: across client engagements we have identified and prioritised 100+ AI opportunities by ROI, and where the resulting redesign has been implemented, it has produced 50%+ cost reduction with service levels improved rather than degraded (Inflection Ventures, delivered engagement results). Those are not projections. They are what has already been measured on operations that were running before the redesign started.
What Each Path Costs to Attempt
Revenue growth in a labour-dependent business is rarely free of the cost base. Inflection Ventures' own diagnostic work across client operations describes the mechanism plainly: to double revenue, a labour-first operator typically has to roughly double staff, so cost grows in step with revenue and margin never improves with scale. That is a structural property of the operating model, not a management failing. It is what "labour-first" means.
Which makes revenue growth, in this kind of business, a capital- and time-intensive bet before it is a margin story. New revenue has to be sold, then delivered, which means hiring ahead of it, training into it, and carrying the ramp-up cost before the revenue lands and often for some time after. None of that is guaranteed to convert. It depends on demand existing, on winning it against competitors, and on execution holding through the ramp. The cost is committed before the outcome is known.
The difference is not that margin expansion is easy. It is that its cost can be scoped in advance, and revenue growth's cannot.
Margin expansion starts from a different position: the revenue is already there. The task is to redesign how the existing work gets done so the same output requires less cost, which is a bounded, scopeable piece of work rather than an open-ended one. It is also, in Inflection Ventures' own delivery data, a lever with a known shape: across client engagements we have identified and prioritised 100+ AI opportunities by ROI, and where the resulting redesign has been implemented, it has produced 50%+ cost reduction with service levels improved rather than degraded (Inflection Ventures, delivered engagement results). Those are not projections. They are what has already been measured on operations that were running before the redesign started.
The difference is not that margin expansion is easy. It is that its cost can be scoped in advance, and revenue growth's cannot.
Why the Two Paths Aren't Equally Available
Why the Two Paths Aren't Equally Available
The scoping difference has a second consequence: who is actually in control of the outcome. Revenue growth in a labour-first business depends on customers, competitors and market timing: three variables no operator sets alone. Margin expansion depends on a decision the operator can make on a Monday morning: redesign this process, remove this handoff, put this workflow through the machine rather than through a person. The work is internal, which means the timeline is too.
The evidence that this internal lever moves the number is not theoretical. McKinsey's study of 471 privately held, PE-backed companies with revenue between $1m and $250m found that businesses that had embedded AI into their operating model, not merely adopted tools opportunistically, ran roughly 20% higher revenue per employee than opportunistic adopters, after accounting for inflation, and traded at a median 14x revenue multiple against 13x for opportunistic adopters (McKinsey, Beyond Productivity: How AI Creates Value in Private Equity, published 23 June 2026). Neither of those companies necessarily grew revenue faster. The operating-model group grew what each unit of revenue was worth.
That one-turn gap is real but capped: McKinsey's higher tiers, where AI reshapes the product or builds new revenue lines entirely, traded at 20x and 31x respectively. Operating-model work buys margin. Product-level reinvention buys the re-rating. The point for a labour-first operator is not that margin expansion is the only lever worth pulling. It is that margin expansion is the one available to pull first, on a timeline the operator controls, without betting the cost base on demand that has not yet arrived.
There is a second, older pattern worth naming here. McKinsey's review of more than 100 buyout deals entered from 2010 and exited by 2021 found that roughly two-thirds of total return came from market multiple expansion and leverage (forces outside an operator's control), while general partners who focused on operational value creation instead achieved returns 2 to 3 percentage points higher on IRR, on average (McKinsey, Bridging Private Equity's Value Creation Gap, published 12 April 2024). The market is providing less of that free lift now than it did across that decade. What is left, structurally, is the lever inside the business.
In a labour-first business, margin expansion is the faster, cheaper and more controllable path to a higher valuation. Revenue growth is the one that requires the market's permission.
Where Operators Get the Choice Backwards
The scoping difference has a second consequence: who is actually in control of the outcome. Revenue growth in a labour-first business depends on customers, competitors and market timing: three variables no operator sets alone. Margin expansion depends on a decision the operator can make on a Monday morning: redesign this process, remove this handoff, put this workflow through the machine rather than through a person. The work is internal, which means the timeline is too.
The evidence that this internal lever moves the number is not theoretical. McKinsey's study of 471 privately held, PE-backed companies with revenue between $1m and $250m found that businesses that had embedded AI into their operating model, not merely adopted tools opportunistically, ran roughly 20% higher revenue per employee than opportunistic adopters, after accounting for inflation, and traded at a median 14x revenue multiple against 13x for opportunistic adopters (McKinsey, Beyond Productivity: How AI Creates Value in Private Equity). Neither of those companies necessarily grew revenue faster. The operating-model group grew what each unit of revenue was worth.
That one-turn gap is real but capped: McKinsey's higher tiers, where AI reshapes the product or builds new revenue lines entirely, traded at 20x and 31x respectively. Operating-model work buys margin. Product-level reinvention buys the re-rating. The point for a labour-first operator is not that margin expansion is the only lever worth pulling. It is that margin expansion is the one available to pull first, on a timeline the operator controls, without betting the cost base on demand that has not yet arrived.
There is a second, older pattern worth naming here. McKinsey's review of more than 100 buyout deals entered from 2010 and exited by 2021 found that roughly two-thirds of total return came from market multiple expansion and leverage (forces outside an operator's control), while general partners who focused on operational value creation instead achieved returns 2 to 3 percentage points higher on IRR, on average (McKinsey, Bridging Private Equity's Value Creation Gap). The market is providing less of that free lift now than it did across that decade. What is left, structurally, is the lever inside the business.
In a labour-first business, margin expansion is the faster, cheaper and more controllable path to a higher valuation. Revenue growth is the one that requires the market's permission.
Where Operators Get the Choice Backwards
From Diagnosis to Value Creation: The Inflection Ventures Approach
Inflection Ventures approaches AI-enabled value creation through two complementary paths:
From Diagnosis to Value Creation: The Inflection Ventures Approach
Inflection Ventures approaches AI-enabled value creation through two complementary paths:
Which Path Can You Actually Pull Today?
Which Path Can You Actually Pull Today?
Revenue growth is not the wrong lever. It is the more expensive one to reach for first in a business where every new dollar of revenue still needs a person behind it.
Margin expansion does not ask the market for anything. It asks the operator to look at work that is already being paid for and decide whether it needs to be done the way it is being done.
That is where Inflection Ventures starts. We look at where the operating model can change, where AI can improve the economics of existing work, and how those changes can translate into measurable P&L impact.
Revenue growth is not the wrong lever. It is the more expensive one to reach for first in a business where every new dollar of revenue still needs a person behind it.
Margin expansion does not ask the market for anything. It asks the operator to look at work that is already being paid for and decide whether it needs to be done the way it is being done.
That is where Inflection Ventures starts. We look at where the operating model can change, where AI can improve the economics of existing work, and how those changes can translate into measurable P&L impact.
If you are exploring where AI could unlock greater efficiency, stronger margins or new enterprise value in your business, talk with the Inflection Ventures team!
If you are exploring where AI could unlock greater efficiency, stronger margins or new enterprise value in your business, talk with the Inflection Ventures team!
Copyright © 2026 Inflection Ventures. All rights reserved.
Copyright © 2026 Inflection Ventures. All rights reserved.