A search fund's edge isn't capital — it's operating attention. In service businesses, AI now compresses variable cost in delivery and fixed cost in overhead at the same time. Both lines move, and they multiply.
A typical mid-market service company runs at gross margins of 30–50% and EBITDA margins of 5–20%. Most of the P&L is labor — the people doing the work, plus the people managing the people doing the work. Capital intensity is low; talent intensity is everything.
That's the structure AI was built to bend.
The point isn't replacing roles wholesale. It's that each role's output ceiling moves up by 2 – 5×, so revenue grows faster than headcount.
Take a typical mid-market service business. Same revenue, same customers, same competitive environment. Apply the two wedges.
| Revenue | 100 |
| Cost of services (variable) | (60) |
| Gross margin | 40% |
| SG&A (fixed) | (25) |
| EBITDA margin | 15% |
| Revenue | 100 |
| Cost of services (variable) | (42) |
| Gross margin | 58% |
| SG&A (fixed) | (18) |
| EBITDA margin | 40% |
Illustrative — not a forecast. Real deployments vary by industry, complexity, and team readiness. The shape, not the precise numbers, is what matters: both lines move, and they multiply.
That's 2.7× EBITDA from operating leverage alone — before any growth. Then apply growth on top, because the same headcount can now serve a larger book.
For a search-fund acquisition, that math is the whole game: buy at 4 – 6× EBITDA, expand margin via AI deployment, exit at a multiple that's both higher (because the business is now scarcer and more profitable) and applied to a larger EBITDA. The compounding is multiplicative, not additive.
Cost compression is the first move because it's the most certain — operating leverage on a fixed cost base. Once that lands, the same AI capability that replaced repetitive cognitive work in delivery can generate revenue against a now-leaner cost base. Growth, with the unit economics already fixed.
AI-driven prospecting and qualification — 5–10× the lead volume at constant headcount, with better targeting because models read context faster than humans triage it.
Tailored outreach, dynamic pricing, and AI-assisted proposals lift close rates without adding sales bodies. Each rep handles a larger book at a higher conversion rate.
Automated KYC, contract drafting, and account setup compress time-to-revenue from weeks to days. Cash conversion accelerates and customers reach value faster.
AI-driven customer success spots churn signals early and surfaces upsell triggers in usage data. Net revenue retention climbs without a proportional CS headcount.
The same AI infrastructure that automates the core service can be repointed at adjacent customer segments or geographies — entry costs that used to require a new ops team are now near-marginal.
Picking up where the first table left off — same business, now with the cost base already restructured. Sales automation drives 30–60% revenue growth over the same 3 – 5 year operating window. The cost lines barely move, so almost all of the incremental revenue drops to EBITDA.
| Revenue | 100 |
| Cost of services (variable) | (42) |
| Gross margin | 58% |
| SG&A (fixed) | (18) |
| EBITDA | 40 (40%) |
| Revenue | 150 |
| Cost of services (variable) | (57) |
| Gross margin | 62% |
| SG&A (fixed) | (18) |
| EBITDA | 75 (50%) |
Variable cost rate compresses one notch further — from 42% to ~38% of revenue — as scale unlocks additional efficiency on top of the AI deployment. SG&A holds flat from Phase 1 in absolute terms: the back-office automation absorbs a 50% larger book at the same fixed-cost dollars. Each incremental dollar of revenue carries a marginal contribution above 60%, so the P&L compounds quickly.
Same business, same customers, same starting point. Across a typical 3 – 5 year operating phase, the two AI wedges deliver the cost-compression first, then the growth — and they compound.
5× EBITDA from operating improvements alone. Then on exit, a business with this profile — high margin, growing 30–50%, AI-native operating moat — typically commands a multiple roughly 2× of what the same business sold for at entry. Acquired at 5× EBITDA, exits at 10×.
Multiplied through: 5× EBITDA expansion × 2× multiple expansion = 10× total enterprise-value creation over the operating phase. That's the search-fund return profile in numerical form, with AI as the operating wedge that makes top-quartile outcomes more reachable than they used to be.
Three things converged in the last 24 months. The mid-market service business that hasn't yet adopted AI is where the asymmetric opportunity lives — larger competitors have started; SMBs haven't. The gap is the alpha.
Frontier models can now do real cognitive work — reading, writing, reasoning, judgement — at near-human quality on domain-specific tasks. The deployable surface in a service business is no longer limited to chatbots.
Inference costs have dropped roughly 10× year over year and continue to fall. Always-on deployment across the workforce — not just pilots — is now economically viable for a sub-$100M-revenue business.
APIs, agents, and fine-tuning are accessible without an in-house ML team. A capable operator-CEO with a small technical partner can deploy what would have required a 20-person research org five years ago.
The same AI capability is available to public companies, PE roll-ups, and SMB owners. The vehicle determines who actually captures the value.
Unlike a minority investor, a search fund principal becomes CEO. AI deployment is directed, not begged for — and it touches every function on the same operating cadence.
AI in operations takes 12 – 24 months to compound into EBITDA. A four-to-ten-year hold lets the value land first in margin, then in exit multiple.
A $5 – 30M EBITDA business has more white space for AI gains than a $500M one — fewer legacy systems to integrate around, simpler decision rights, faster deployment cycles.
Not every AI deployment moves margin. The 30% of activities that consume 70% of low-margin labor are where we focus first.
If you're an investor interested in the asset class, an owner of a service business approaching transition, or an operator who wants to compare notes — we're listening.