A Personal Point of View on AI-Driven Value Creation in Private Equity
By Ronny Fehling, Chief AI Transformation Officer, HTEC
Private equity does not suffer from a lack of ideas.
Across diligence, investment committee discussions, and the first hundred days, value creation plans are typically well articulated. Upside cases are modeled, AI opportunities are identified, and transformation roadmaps are drawn. Yet despite this apparent clarity, a large share of these initiatives fail to materialize in a way that meaningfully moves EBITDA or holds through exit.
The problem is not insufficient ambition.
It is a widening gap between conviction and execution reality — a gap that artificial intelligence has made larger, not smaller.
HTEC exists to close that gap.
The Core Problem: AI Has Changed the Nature of Execution Risk
AI has fundamentally altered how value can be created. It has also changed how easily value creation can stall.
Many organizations now claim to “use AI,” yet in practice that usage usually falls into one of three categories:
- AI as a marketing claim
- AI as an isolated feature
- AI as a true economic differentiator embedded into operations
Only the third reliably creates defensible value at exit.
For private equity, the challenge is that these distinctions are often invisible at diligence or early planning stages. Slideware looks convincing. Architectures look plausible. Business cases look attractive. But once execution begins, hidden constraints surface: data reality, technical debt, organizational friction, regulatory exposure, and simple underestimation of what it takes to move from concept into production.
This is where value creation plans lose time.
And in private equity, time is the most unforgiving constraint of all.
HTEC’s Core Belief
Private equity does not struggle to identify value.
It struggles to turn conviction into production-grade outcomes that move EBITDA and compound across the hold period — especially in AI-driven transformations.
The point of view that guides my work, and the way we operate at HTEC, is simple and explicit:
AI value creation only matters if it reaches production, survives operational reality, and compounds economically over time.
Everything else is noise.
The HTEC Lens: Four Layers, One System
HTEC approaches AI-driven value creation in private equity as a single, integrated system built on four layers. Each layer is necessary. None is sufficient on its own.
1. Investment Logic: Why Value Must Compound
AI initiatives must be evaluated through a private-equity-native lens:
- Value must compound across the investment cycle, not peak early and decay
- Initiatives must survive leadership changes, organizational friction, and ownership transitions
- Exit multiple expansion matters more than isolated operational wins
This means prioritizing initiatives that build durable capabilities—technology, data, AI, and operating models—rather than one-off optimizations that disappear before exit.
This layer defines what matters.
2. AI Value Lever Discipline: What Is Worth Doing
Not all AI initiatives are equal. Most should never be funded.
HTEC applies a strict value-lever taxonomy to filter signal from noise. AI is only pursued where it can realistically move levers that matter to private equity:
- Cost and margin expansion
- Productivity and throughput
- Revenue growth and pricing power
- Speed to market
- Risk reduction and operational resilience
- Strategic optionality at exit
This is not a catalog of use cases.
It is a decision filter.
If an AI initiative cannot be credibly linked to one or more of these levers, it does not proceed.
This layer defines what is economically relevant.
3. Execution Reality and Agentic Systems: How Value Reaches Production
Execution is where most AI strategies fail.
HTEC operates with a production-first mindset:
- No pilot theatre
- No long-running initiatives without economic proof
- No architectures that cannot survive real-world constraints
Agentic systems are used not as a buzzword, but as an execution mechanism — reducing human friction, automating repeatable decision flows, and enabling scale without linear cost growth. Human-in-the-loop models are applied deliberately where economics, safety, or regulation require it.
Crucially, systems are designed to work with the organization as it is — not as it might ideally become.
This layer defines how value becomes real.
4. Decision Discipline: Where We Differentiate
HTEC’s true differentiation is not technology.
It is decision discipline under uncertainty.
Every initiative is governed by explicit ROI thresholds and stop/go decisions:
- Scale when value materializes
- Reset when assumptions change
- Kill early when the economics do not hold
HTEC embeds forward-deployed teams directly into portfolio companies, operating close to reality and close to decision-makers. If EBITDA does not move, the approach changes—or stops.
This discipline protects the most expensive asset in private equity: time – losing a year of the hold period on initiatives that never scale.
It also ensures that value creation compounds rather than stalls.
This layer defines why value creation compounds instead of stalling.
Why “Use-Case POCs” Fail in Private Equity
A common response to AI uncertainty in PE is to ask for prioritised AI use cases, launch multiple POCs in parallel, and then “see what sticks.”
This approach feels fast. In reality, it is one of the most reliable ways to lose a year of the hold period.
The failure mode is structural. Starting from use cases rather than value levers and execution feasibility produces impressive demonstrations — and very few systems that survive contact with production.
AI value in private equity rarely fails because ideas were wrong.
It fails because execution assumptions were optimistic.
A Different Definition of POC: The First Production Slice
This is where the concept of a “POC” becomes dangerous.
In many organizations, a POC is treated as a demo: limited data, simplified assumptions, no integration into real workflows, and no credible path to scale. These POCs reduce technical uncertainty but leave economic and operational risk untouched.
In private equity, that is not sufficient.
A meaningful POC must be treated as the first production slice:
- It uses real or production-adjacent data
- It runs in the actual operating environment
- It reflects real constraints — security, governance, latency, cost
- It is explicitly designed either to scale or to be killed quickly
The purpose of a POC is not to prove that something is possible.
It is to prove that something is worth scaling.
Anything else is theatre.
Where HTEC Engages in the Investment Cycle
HTEC engages where execution reality matters most.
Diligence and Early Conviction
During diligence or immediately post-close, the most valuable contribution is not another assessment of potential. It is clarity on execution reality:
- Assessing true technology and AI readiness at the production level
- Is AI merely claimed, or embedded as a real differentiator?
- Which value levers can realistically be moved within the hold period?
- What will it actually take — in time, capital, and organizational effort — to reach production?
Assessing true technology and AI readiness at the production level
Grounded estimates, informed by execution experience, are often more valuable than optimistic ones. And because HTEC builds and runs systems in production, these assessments are based on lived execution experience—not optimistic assumptions.
The outcome is conviction grounded in reality, not aspiration.
Strategic Reality Check: Defense, Offense, Right to Exist
In parallel, external reality must be assessed:
- How AI is reshaping the value chain
- Which capabilities are now non-optional
- Where the asset has credible offensive differentiation
- What must be true for the equity story to hold at exit
This avoids backing assets that look attractive today but are structurally exposed tomorrow.
Post-Close Execution and Compounding
Once conviction exists, execution must follow:
- Moving priority initiatives into production
- Enforcing decision and ROI discipline
- Building capabilities that compound across the hold period
The goal is not activity.
The goal is durable value that survives to exit.
Red Ocean and Blue Ocean in AI Value Creation
Much of the market today operates in a red ocean:
- Identifying AI opportunities
- Framing value narratives
- Producing prioritisation lists and POCs
These activities are important—but increasingly table stakes.
The blue ocean lies elsewhere:
- Deciding which bets are worth making
- Turning conviction into production under real constraints
- Compounding value across the hold period — or stopping early
That is where AI becomes an economic lever rather than a distraction.
A Final Point of View
AI does not create value because it is advanced.
It creates value because it is executed with discipline.
For private equity, the question is no longer whether AI can be useful.
The question is whether it can be turned into production-grade, EBITDA-relevant outcomes fast enough to matter — and robust enough to hold at exit.
Everything else is noise.
