The biggest mistake PE firms make with AI is the "one-and-done" approach — writing off the technology after a single bad experience prevents teams from building the iterative skillset that compounds over time, and firms willing to experiment are quietly pulling ahead of those waiting for a perfect answer

General AI tools like ChatGPT Enterprise are insufficient for PE workflows — real diligence involves tens of millions of tokens across hundreds of unstructured documents that don't talk to each other, requiring purpose-built document ingestion, vector-based knowledge structures, and fully segregated data environments with NDAs

Mid-market and lower mid-market firms are actually adopting AI faster than mega-funds — large caps are slowed by organizational inertia and the ambition to build proprietary infrastructure, while smaller teams see a 2–3x productivity multiplier on associates as an existential competitive necessity rather than a nice-to-have

AI will flatten the PE org chart by 2027, but it raises the bar for human judgment rather than lowering it — associates will cover more ground and headcount won't scale proportionally with deal volume, making the ability to push back on model output and take an adversarial analytical position more valuable, not less

LPs are beginning to ask about AI as part of fund diligence — questions around how AI is being used in sourcing, diligence, and portfolio support are becoming standard in annual meetings, signaling that AI capability may soon be a baseline expectation rather than a differentiator

Having an MBA is the single most impactful controllable demographic factor in ETA — MBA-led searches have a median IRR of 28% vs 1% for non-MBAs and a mean IRR of 22% vs -28%, with non-MBAs underperforming across every metric and suffering dramatically more total losses, a gap that remains statistically significant at the 5% level even after controlling for other variables

Geography is the strongest structural predictor of IRR, with US/Canada funds outperforming international peers by ~36 percentage points — outside North America, median IRR drops to 9% vs 28% and median MOIC falls to 1.7x vs 3.0x, likely reflecting weaker debt markets, thinner ETA ecosystems, and less developed exit infrastructure

MBA prestige matters far less than simply having an MBA — median IRRs are broadly similar across top-15, M7, and non-M7 programs, with Kellogg leading at 46% median IRR but from only 9 observations, and Harvard/Stanford actually trailing other MBAs on median IRR (22% vs 30%) despite their dominant brand reputation in ETA circles

Partnerships appear to outperform in single-variable analysis but the advantage largely disappears in multivariable regression — partnered searches show a median IRR of 32% vs 24% for solo, but once education and geography are controlled for, the partnership premium mostly vanishes, suggesting the benefit reflects correlated traits rather than teaming up itself

Observable demographics collectively explain only ~26% of IRR variation, with the rest likely driven by unobservable traits — professional backgrounds in banking, PE, and consulting show directional strength (median IRR 29% vs 22%), military veterans outperform on IRR (median 35% vs 24%), and women perform on par with men, but none of these factors are statistically significant in the multivariable model, pointing to grit, coachability, and leadership as the true differentiators

Operating leverage occurs when revenue grows faster than overhead, causing net profit to expand disproportionately — in the example, a 15% revenue increase from $33K to $38K/month paired with only a 9% overhead increase drives net profit up 27%, from $5.5K to $7K, while gross margin stays flat at 50%

The key distinction is between gross margin and net margin — gross margin reflects the efficiency of serving each dollar of revenue and stays constant in the example at 50%, while net margin expands as the fixed overhead burden gets spread across a larger revenue base, rising from 17% to 18.4%

Overhead growth rate is the critical variable to watch — when overhead scales at the same rate as revenue there is zero operating leverage and margins stay flat, but when overhead growth lags revenue growth even modestly, every incremental dollar of revenue drops to the bottom line at an accelerating rate

Small businesses unlock real profitability not by raising gross margins but by outgrowing their cost base — the moment a business can add meaningful revenue without proportionally adding overhead is when net profit starts compounding faster than topline growth, which is the structural shift that separates scaling businesses from ones that simply stay busy

Operating leverage is most powerful when understood as a design principle, not a byproduct — deliberately building systems, processes, and infrastructure that can support higher revenue without requiring proportional overhead additions is what allows small business owners to convert growth into wealth rather than just more complexity