by Priya Raman | GOTARA | AI, Artificial Intelligence, PMI, TRANSFORMATION
For the first time in decades, powerful technology is available without massive upfront capital. The access gap that defined every prior technology wave has narrowed. Mid-market companies can now reach AI capabilities that once required Fortune 500 infrastructure budgets. That is a genuine opportunity. It is also why every board deck is full of AI—and why the investment committee (IC) is still not hearing a number.
TLDR;
AI has eliminated the access gap. The problem today is value translation.
The fastest path to AI value starts by working backward from the customer moment that matters: where customers churn, waiting, abandonment, under-buying, complaining, or failing to get value.
Most companies know where to deploy AI. Very few have identified what is blocking the value—data gaps, broken processes, missing ownership, legal constraints, skill gaps, or unclear decision rights.
The organizations closing the gap are connecting customer value moments to P&L outcomes, operating model changes, and accountable owners.
There is a specific set of moves that separates portfolio companies generating measurable AI value from ones running expensive pilots.
AI Is Different. The Value Gap Still Exists. Here Is Why.
I have seen this firsthand as a senior Data and AI executive across Fortune 500 and PE-backed environments, leading to commercial and operating transformations where value expectations are real, capital is scrutinized, and the pressure to perform is unambiguous.
Every prior technology wave created an access gap. Large enterprises could afford the platforms, infrastructure, systems integrators, and multi-year transformation programs. Small-to-medium companies often could not. The deepest pockets got the earliest advantage. I watched this play out across three decades of technology cycles. The pattern was consistent: the firms that funded the platform and not the execution model spent years and did not see their numbers move.
AI has changed the access equation. For the first time, powerful technology is available without the same level of upfront capital. A mid-market portfolio company can access capabilities that once required massive infrastructure, specialist teams, and years of build time. The access barrier has dropped dramatically.
However, the shift has made the market noisier, and AI is increasingly treated as the magic pill for every problem. AI investment is accelerating. Board decks are full of AI strategy. Every function has ideas. Tools are everywhere. Demos are impressive. And when the IC asks what AI has done for EBITDA this quarter—or which metric has moved on customers, revenue, margin, or cost-to-serve—the answer is rarely a number.
The problem is not lack of access. The problem is lack of value translation. Most companies are not failing because they lack AI tools or leadership motivation. They are failing at the last mile. What has worked instead—starting with a customer or business outcome, working backward to the decision or workflow that must change, and converting AI capability into financial impact.
What Is Actually Blocking the Value
In my operator experience across Fortune 500 companies and PE-backed portfolios, the blockers to AI value creation are remarkably consistent. They do not show up the same way in every use case. Still, the pattern is familiar: data quality, accuracy, trust, permissions, semantics, process readiness, ownership, adoption, change management, capabilities, and skills.
In most organizations, several of these blockers are present at the same time, but not at the same intensity for every opportunity. That is why the AI value gap persists even in companies with strong technology, serious investment, and capable leadership.
Most AI programs still start with use cases, not customer or business outcomes. Without a clear line to the customer friction point and the P&L metric it affects, effort disperses. Teams build capability, launch pilots, train users, and deploy tools, but the work does not reliably convert into measurable enterprise value.
No clarity on where AI creates measurable enterprise value. The question that should precede every AI investment is: which customer problem, operating decision, or business number will improve as a result, what process will change, and who is accountable for it? Most programs have never answered this question clearly.
Undiagnosed blockers between AI and value. For example: Data gaps that were never mapped. Broken processes that AI now accelerates. Legal and compliance constraints on what data can be used. Skill gaps in the teams meant to act on AI outputs. Leadership ambiguity on who decides and who owns the result. Most organizations have several of these. Very few have diagnosed all of them.
No ownership and adoption strategy for the last mile. The last mile is the conversion of AI output into a customer-facing decision, functioning workflow change, or measurable business outcome. It requires the same discipline as any other scaled change program: a named owner, a defined action, a tracked metric, and an incentive structure that rewards the outcome. Despite best intent, most AI projects skip it entirely.
Progress is measured in activity, not financial impact. For example, when pilots are launched, users are trained, models are deployed, it feels like momentum. They are not valuable until a P&L line moves and someone’s name is on it.
The blockers between AI investment and business impact are specific, diagnosable, and not fixed by adding more tools, tech, consultants, or models. The value is found by working backward from the customer and business outcome, then diagnosing what prevents the organization from acting on the signal. That requires operator judgment.
Two Use Cases. Same Root Problem. Very Different Outcomes.
The two examples below, drawn from two different industries, highlight different root causes and different paths to value. In each case, value was realized only when the organization worked backward from a customer or business outcome and connected the capability to the right operating model, ownership, adoption path, and last-mile execution.
The outcomes are real. What produced them was not the model alone. It was the full system required to convert AI capability into a better customer moment, a better decision, and measurable enterprise value.
This is where operator depth matters. The work is not to boil the ocean, but to know where to look first. Experienced operators can see where customer friction, decision stalls, ownership breaks, adoption fails, incentives conflict, and value gets lost between strategy and action. That is the last mile of AI value creation, where measurable outcomes are either captured or missed.
What Separates Value-Creating AI Programs From Expensive Pilots
I have seen portfolio companies close the gap in 90 days. I have seen others spend years without producing a single outcome metric they could defend to the IC. The difference is never the technology, never the vendor, and rarely the management team’s willingness to act.
It comes down to five specific moves. Each addresses a different constraint that prevents AI from translating into measurable business value. Every portfolio company encounters some version of these constraints. The difference is not whether they exist. The difference is whether leadership identifies the right constraint, addresses it in the right sequence, and connects it to an outcome the board can measure.
These are the top two moves we recommend:
Start with the customer value moment and the exit metric, not the use case.
- Most teams start with the technology. The ones creating value start with the customer friction point, the business number, and the exit metric, then work backward. There is a specific diagnostic for this.
Fix the decision architecture, not just the data pipeline.
- AI amplifies the quality of the decision process it sits behind. In most portfolio companies, that process was designed before AI existed. What needs to change is not obvious—and it is different in every business.
The Question Most Operating Partners Have Not Yet Asked
Before your next review, ask your management team one question. Not about the roadmap. Not about the vendor. Not about adoption curves or models in production.
Which customer, revenue, margin, cost-to-serve, or cycle-time metric has measurably improved in the last 90 days as a direct result of AI—and who is accountable for it?
Not a pilot result. Not a projected saving. Not a productivity estimate. A number already in the management accounts. Revenue per customer. Gross margin. Churn rate. Cost-to-serve. Cycle time to close. Customer NPS. Renewal rate.
If the answer is a number with a name attached—your program is creating value. If the answer is a roadmap, a list of initiatives, or a progress update—the gap is open and compounding.
In my experience, fewer than one in five PE-backed companies running active AI programs can answer yes to that question today. That is not a failure of technology or ambition. It is a structural problem in the way most AI programs are designed—and it sits in a specific place that a standard AI maturity assessment will not find.
The Value Is There. The Question Is Whether You Know Where to Look.
The gap sits in a specific place most AI reviews do not reach. It is not always obvious from the model, the platform, or the use case list. It shows up in the space between capability and value: ownership, decision rights, adoption, incentives, workflow change, governance, and execution.
Closing that gap is not complicated once it is correctly diagnosed. But it requires a different conversation, one that starts with the customer and business outcome and works backward, not one that starts with the technology and works forward.
The companies that will exit with AI-driven multiple expansion will not necessarily be the ones with the most sophisticated models. They will be the ones that identified the value conversion gap early, before the hold period compressed, before the competitive window narrowed, and before the board started asking why the numbers had not moved.
Most of those companies will not start with a bigger budget. They will start with a better question, and the right operator beside them to answer it.
What PE Firms Ask Me Most (FAQ)
Q: How do I know if my portfolio company has an AI value gap?
A: Ask one question at the next board review: which customer, revenue, margin, cost-to-serve, or cycle-time metric has measurably improved in the last 90 days as a direct result of AI, and who owns it? If the answer is not a number, the gap is open. In my experience, it is open in the majority of active AI programs—including ones with sophisticated technology and capable management teams.
Q: Can this be fixed within a typical hold period?
A: Yes – and faster than most people expect, once the right diagnosis is in place. I have seen portfolio companies move from zero measurable AI value to meaningful EBITDA contribution in under 90 days. The fix is not a technology upgrade or a new vendor. It requires a specific set of customer-back operating model decisions. The reason most management teams have not made them is that nobody has asked for them in the right way.
Q: What is the most common mistake Operating Partners make with portfolio AI programmes?
A: Approving the technology budget without asking what customer outcome, operating model change, and P&L result are required to realize the value. When technology comes first, and execution comes later, the gap opens in the space between them. And it compounds quietly until the next board review makes it visible.
Continue Reading
Looking for more insights on AI, Private Equity, Business Transformation, and M&A?
Top comments (0)