PE firms have what no other owner has: control, concentration, and a value creation mandate. One firm, one playbook, applied across an entire portfolio. What most mid-market firms lack is the capability layer that turns AI conviction into measured EBITDA impact. That layer is the work we do.
The era when leverage and rising multiples carried returns is over. Value now has to be built inside portfolio companies, and AI is the largest operational improvement lever available. The industry knows it. Execution is another matter.
The leading firms have moved decisively. EQT runs a dedicated AI platform across its investment lifecycle. Blackstone's data science team has worked inside more than seventy portfolio companies. Vista reports 80 per cent of its majority-owned companies deploying generative AI, with coding productivity gains of up to 30 per cent. The frontier AI labs have noticed too: Anthropic and OpenAI have each launched billion-dollar ventures aimed at deploying AI across PE-owned companies.
Yet the results below the headline tier tell a different story. Around half of general partners report that AI initiatives in their portfolio companies are falling short of expectations. The reasons are rarely technical. Portfolio companies are mid-market businesses, and they fail at AI adoption for mid-market reasons: unclear value pools, no internal AI capability, weak data foundations, and tools bought but never properly worked into how people operate. An owner's mandate gets the licences purchased. It does not get them adopted.
For mid-market and lower mid-market firms the gap is structural. The mega-funds have in-house data science teams and direct lab partnerships. A £200m to £1bn fund has the same motive and none of that machinery: a lean deal team, a small or non-existent operating bench, and ten to twenty portfolio companies that each need what a Vista or Blackstone company gets from its owner.
There is also a sharpening external pressure. LPs increasingly ask how GPs are using AI across investment and portfolio operations, and buyers now assess AI readiness during diligence. AI capability is becoming part of both the fundraising story and the exit story.
Mike led enterprise-wide AI transformation at Verimatrix, a publicly listed global SaaS business of exactly the scale and shape of a mid-market portfolio company, under direct ExCom oversight. Not advising on adoption. Executing it, function by function, across nine countries.
That programme covered the full range an operating partner cares about: engineering productivity with IP protections, AI-assisted sales workflows that moved win rates, support automation that cut ticket volumes, and a Responsible AI governance framework approved at board level. The outcomes were measured, attributed and reported to a listed-company executive committee, the same standard of evidence a value creation plan demands.
Human-AI Systems itself runs on the model we build for clients: a team of ten AI agents handling operations, finance, marketing, research and delivery support, directed by one experienced operator. We operate the operating model we recommend.
AI lands differently at each stage of the deal cycle. The compounding returns come from treating it as firm-level infrastructure rather than a series of one-off projects.
AI-assisted market analysis, data room interrogation and risk scanning are now standard practice in deal teams. The newer question is the reverse one: assessing the AI readiness and AI exposure of the target itself. What does AI do to this company's market? Where is the unrealised productivity? That assessment belongs in diligence, because it prices the value creation plan.
A PE firm is a lean knowledge-work business: deal memos, monitoring packs, LP reporting, fundraising DDQs, internal research. It is often the right first deployment, because a GP that has experienced the change personally sponsors portfolio adoption with far more conviction than one working from a slide.
The first 100 days install the reporting pack, the governance and the value creation plan. AI adoption belongs in that same window: the firm's AI standard, tools, training and playbook deployed as part of integration, not discovered ad hoc in year two. For buy-and-build strategies this repeats with every bolt-on.
The core value creation work: redesigning workflows in sales, service, engineering, finance and operations so AI moves the numbers that matter at exit. This is where around half of initiatives currently disappoint, and almost always for adoption reasons rather than technology reasons. Structured assessment, piloting in live workflows and disciplined scaling is the difference.
The highest-leverage move in the whole lifecycle: codify what works in one company into a shared library of AI skills, agents, prompts, training and operating processes that every portfolio company pulls from and contributes back to. Built once, deployed many times, and retained by the firm as an asset that survives any individual exit. The leading firms run exactly this model. Very few below the mega-fund tier do.
Buyers now diligence AI claims technically. A portfolio company with a documented AI operating model, measured adoption and attributable financial impact tells a stronger equity story and survives a sharper process. With exit backlogs at record levels, making the AI story diligence-proof is becoming part of preparing any asset for sale.
One relationship, portfolio-wide reach. Engagements are structured so that what is learned in one company compounds across all of them.
Whether the starting point is a single underperforming AI initiative, a new platform acquisition, or a firm-level ambition to build a portfolio AI capability, the approach is the same: assess honestly, pilot in live workflows, scale what is proven, and codify it so the firm keeps the asset.
A structured assessment across the portfolio and the firm itself. Maps each company against our Tool, Assistant, Worker capability model, identifies the two or three companies and use cases with the best value-to-effort ratio, and gives the firm an evidence-based deployment sequence instead of a wish list. The natural first engagement.
Our core Assess, Pilot, Scale engagement delivered inside individual portfolio companies, with the owner as sponsor. Workflow redesign, measured pilots in live operations, governance that holds up at board level, and outcomes expressed in the language of the value creation plan: revenue growth, margin expansion, and capability the next owner will pay for.
Building the firm's shared AI capability: the skills library, the playbooks, the training, the governance standard, and the mechanics for portfolio companies to contribute improvements back. Includes embedding AI adoption into the firm's onboarding standard so every new acquisition starts from the playbook rather than from zero.
AI readiness assessment of targets as a defined diligence workstream, and AI-readiness work ahead of exit: documenting the operating model, evidencing adoption and attributing impact so the AI story strengthens the equity story rather than inviting a discount.
A retained engagement, typically one to three days per month, providing the AI operating capability the largest funds hire full-time. Sitting alongside deal partners and management teams: reviewing portfolio AI progress, unblocking stalled initiatives, advising on AI questions in live deals, and steering the standardisation programme. Senior enough for the investment committee, hands-on enough for a management team.
Structured AI value creation for mid-market PE, from an operator who has delivered the transformation inside a portfolio-scale software business and runs his own firm on the model.