About this Q&A: Alex Dukic, Chief Digital Officer at HTEC, makes the case that artificial intelligence has become a capital allocation question rather than a technology one. In this interview, he examines why most PE portfolios remain underexposed to AI risk, what production-grade deployment actually looks like inside a five-year hold period, and where the most overlooked value creation opportunity in private equity currently sits.
Q: What is the most significant shift you are seeing in how private equity firms think about AI across their portfolios right now?
Alex Dukic: The conversation has moved on considerably, though unevenly across the market. Twelve months ago, the dominant question was about cost: how to reduce operational spend, improve margins, and compress the bottom line. That framing still matters, but the firms genuinely ahead have moved past it. The main question now is growth: how do AI capabilities expand revenue, open adjacent markets, improve retention, and build the kind of data assets that a strategic acquirer will pay a premium for at exit?
Buyers are already reflecting this shift in diligence: cost reduction has a ceiling. They can model it, discount it, and price it at entry. A portfolio company that can demonstrate revenue growth from AI already working in production, a live system generating measurable impact rather than a roadmap or just another strategy document, commands a fundamentally different exit valuation. The firms that understand this, while they still have quarters left to act, are building that story now. The ones who leave it too late find the gap exposed when the buyer’s diligence team arrives.
We are also seeing a clear hiring signal at the fund level. PE firms in the US are beginning to appoint Chief AI Officers to look across their entire portfolio, gauging the threat of AI-native competition and the growth opportunity AI represents for their non-tech holdings; that tells you something important about where this has landed strategically. AI has become a fund-level question, and the firms treating it as one are already pulling ahead.
Q: You have argued that AI has become a capital risk rather than a technology risk. Why does that distinction matter for GPs?
Alex Dukic: Most PE firms are excellent at pricing financial risk. They have rigorous frameworks for leverage, market exposure, management quality, and competitive positioning. AI risk is a different kind of exposure, and most portfolios have yet to be stress-tested for it.
The distinction matters because AI is affecting enterprise value on both sides of the equation at once. On the EBITDA side, AI-native competitors are compressing margins and accelerating customer churn in ways that traditional competitive analysis fails to capture quickly enough. On the multiple side, strategic acquirers are assigning a premium to embedded AI capability and a discount to legacy architecture, regardless of how clean the financials look. A company with strong EBITDA margins and no credible AI capability in its core workflows carries more risk than it appears on paper, because buyers will identify that gap even before the GP does.
One portfolio company had a solid exit thesis built on continued margin expansion. In diligence, the buyer’s team identified three AI-native competitors that had entered the market in the prior 18 months. The multiple was renegotiated downward. The information had been available to act on for over a year; by the time it surfaced in diligence, the window had closed.
Q: Where do you see the most underexplored AI opportunity across PE portfolios, and why is it being overlooked?
Alex Dukic: The real opportunity sits in the services businesses, the industrials, the healthcare and professional services firms that AI has barely touched. These are companies with meaningful operational scale, genuine complexity, and an AI capability that is close to zero. That combination, significant upside combined with minimal existing deployment and lower competition for specialist talent, is where the return on AI investment tends to be most asymmetric.
Take a nationwide custom closet manufacturer with genuine operational complexity running from marketing through sales, manufacturing, and installation, and almost no AI applied to any of it. A short execution sprint with leadership surfaced two things: smaller deals were producing little or no profit, and AI-driven matching of prospects to delivery capacity was the highest-return fix. The first initiative was live in three months and returned twelve times the investment in year one. That is the asymmetry: a business nobody would describe as a technology story, generating a first-deployment return that transformation programs at far larger companies rarely deliver.
These businesses are overlooked partly because of expectations. When a portfolio company operates in a historically low-tech vertical, the assumption is that the transformation ceiling is lower. In reality, a business applying AI to its core workflows for the first time has more to gain from a well-executed first deployment than a tech-native business optimizing at the margin. The proof point only needs to be real, measurable, and visible to the board and, therefore, the buyer.
Mid-market businesses have enough scale to make AI economically meaningful, but they have often been deprioritized in favor of larger and more obvious transformation targets. The value creation potential is greatest precisely where the competition for it is thinnest.
Q: Why are so many AI initiatives inside PE-backed companies failing to generate returns, and what does a production-grade deployment actually look like inside a five-year hold period?
Alex Dukic: The root cause is a methodology mismatch. Proof-of-concept culture was designed for enterprise innovation labs, environments where experimentation has its own budget, timeline, and a degree of tolerance for inconclusive outcomes. A four-to-five-year hold period, measured in quarters, with LPs watching every reporting cycle, cannot absorb AI initiatives that spend 18 months in pilot and never reach production.
An 18-month pilot with no production outcome is value destruction, full stop. The incentive structure governing most technology engagements makes this worse. Most AI partners get paid whether or not the initiative reaches production. That misalignment is the structural cause of the POC problem, and outcome-aligned commercial models are the direct solution to it.
The right approach applies the same capital discipline to AI that PE firms bring to every other investment: defined value levers, clear ownership, short horizons, and measurable outcomes. Success means a working system in production, generating measurable business impact, visible in the operating review.
We worked with a PE-backed company in the aesthetic healthcare space, a business acquiring surgery centers across the United States. After-hours patient inquiry volume was significant, coverage was zero, and the revenue impact was understood but unaddressed. A one-day workshop with the executive team identified the opportunity and committed to a production path. A live AI deployment handling real patient interactions and generating real revenue was up and running within six months of that initial conversation. That sequence is the model rather than a happy accident: a short, structured sprint that ends in a production commitment with an owner, a budget, and a measurable target, rather than a report. Two or three programs of that kind, running in production and driving measurable impact, is what success looks like at 12 to 18 months.
Most firms fall short of that benchmark because they optimize for the wrong thing at the outset.
Q: What are buyers actually asking about AI in diligence today, and how prepared are most sellers to answer?
Alex Dukic: Buyers are asking questions most sellers have not prepared for: which workflows have been transformed by AI? What is the revenue or cost impact, and how is it being measured? What is the competitive exposure from AI-native entrants in this market? Does the technology architecture support AI at scale, or constrain growth?
These are now standard diligence items at the more sophisticated acquirers, and they are moving downstream quickly. Most PE-backed companies still need to prepare a coherent answer, because the GP has not required one. The exit narrative was built on financial performance, geographic expansion, or bolt-on M&A. That playbook is already priced in at entry. AI is the remaining lever, and a company that can demonstrate it credibly commands a meaningfully stronger position at exit.
The best acquiring firms are already applying AI to earnings calls, regulatory filings, and operational data to identify targets before they come to market. The information edge that came from relationships and market access alone is narrowing. GPs who are building equivalent data-driven sourcing capability now are compounding an advantage that will be difficult for later movers to close.
Q: HTEC is itself PE-backed. How does that shape the way you work with GPs and portfolio companies?
Alex Dukic: It changes the conversation entirely. We advise from the inside, having lived the same pressures: what it takes to drive a successful exit, how to move a business from a technology and operational perspective under real time and capital constraints, and what it feels like when the LP is watching every reporting cycle. That credibility matters when you are sitting across the table from an operating partner who has seen a great deal of technology consulting that failed to translate into results.
It also shapes our commercial model. We align with investors and portfolio companies on outcomes rather than on delivery alone. In some engagements, that means tying our success directly to the business results we generate. In others, where we engage early in the investment cycle, we take a longer view on how our contribution is recognized at exit. When both sides only succeed if the AI is live and delivering real results, there is no incentive for either party to stay stuck in endless pilot phases, so they don’t.
We also cover the full investment cycle, from technology due diligence before acquisition through value creation planning in the first 100 days, execution and M&A integration, to exit preparation. Many firms that operate in this space complete the diligence and step back, but we stay in. And that continuity is where a significant amount of value is either captured or lost.
Q: Looking ahead, what developments should PE firms be preparing for that have yet to fully register on their radar?
Alex Dukic: The hold period itself is under pressure, and the industry has yet to fully price that in. Aggressive AI deployment is beginning to compress the timeline between acquisition and exit-readiness. What has historically taken four to five years may become achievable in 18 to 24 months for the firms that move early and with genuine conviction. We’re talking about faster capital velocity, stronger returns, and a new generation of exit-ready assets coming to market sooner than the current model anticipates.
The other development worth watching is how AI is reshaping deal sourcing and competitive intelligence. The firms generating proprietary insight from data, applying analysis to filings, earnings calls, workforce signals, and operational indicators before companies come to market, are building a sourcing advantage that is structural rather than cyclical. This capability is being built right now by the most forward-thinking GPs, and the gap between them and the rest of the market will widen quickly.
The PE firms that bring to AI the same rigor and capital discipline they apply to every other investment decision will outperform those that treat it as a technology initiative and measure it accordingly. The framing chosen at the beginning determines almost everything that follows.
