Your AI Strategy Is Not a Strategy — It Is an Expense Account
Somewhere in your organization right now, someone is preparing a presentation about AI. It may be a vendor pitching a large language model integration. It may be an internal champion advocating for an automation pilot. It may be a board member asking why your company has not yet announced an AI initiative, given that every competitor seems to have one.
The pressure to act is real, and it is accelerating. But pressure, in the absence of strategic clarity, produces expensive mistakes — and the AI landscape is currently generating those mistakes at an impressive scale.
This is not a contrarian argument against artificial intelligence. The technology is genuinely powerful, and its capacity to create business value is not hypothetical. It is an argument against the particular form of organizational behavior that the AI hype cycle reliably produces: the rush to implement before the work of thinking has been done.
What "AI-Driven Efficiency" Usually Means in Practice
When a company announces that it is pursuing AI-driven efficiency, the statement typically describes a technology acquisition decision, not a strategic one. A platform has been selected. A vendor has been contracted. A pilot has been launched in a function that volunteered for it. Leadership has been briefed on the use case, and the slide deck includes a projected efficiency gain that the implementation team is now under pressure to validate.
What is usually absent from this picture is a rigorous answer to the foundational question: efficiency in service of what?
Efficiency is not an end in itself. It is a means of freeing resources — time, capital, human capacity — for deployment toward higher-value activities. If an organization automates a process but has no defined plan for how the released capacity will be redeployed, the efficiency gain does not produce a business outcome. It produces a cost reduction that is frequently offset by the implementation cost, the change management burden, and the ongoing licensing fees of the AI system itself.
A 2023 survey by KPMG found that fewer than 30 percent of US companies that had deployed AI solutions could demonstrate a clear, quantified return on investment from those deployments. That finding is not an indictment of AI. It is a description of what happens when organizations treat a strategic problem as a procurement problem.
The Change Management Debt
There is a particular pattern that emerges repeatedly in AI implementations that underdeliver. The technology works as advertised. The technical implementation is completed on schedule. Adoption rates are adequate. And yet, six months after go-live, the business outcomes that justified the investment have not materialized.
In the majority of these cases, the explanation is organizational rather than technical. AI systems change how work gets done. They shift the skills required to perform certain functions. They alter the role of human judgment in processes that were previously fully human-executed. These are not trivial adjustments. They require deliberate change management — communication, training, role redesign, and in some cases, structural reorganization.
Organizations that treat AI implementation as a technology project rather than an organizational change initiative consistently underestimate this requirement. The result is a workforce that is nominally using the new system while working around it in ways that preserve familiar habits and limit the technology's actual impact.
Change management is not a soft discipline. In the context of AI implementation, it is the primary variable determining whether the investment generates returns. Companies that allocate serious resources to it — proportional to the scale of behavioral change the implementation requires — consistently outperform those that do not.
The Strategic Alignment Test
Before any organization commits capital to an AI initiative, it should be able to answer five questions with specificity:
First: What specific business problem does this initiative address, and what is the current cost of that problem in measurable terms?
Second: What outcome will success look like, and what metrics will confirm that the outcome has been achieved?
Third: What is the realistic timeline for the investment to generate returns that exceed its total cost of ownership, including implementation, licensing, training, and change management?
Fourth: What organizational capabilities — skills, processes, governance structures — must be in place for the technology to function as intended?
Fifth: What is the plan for redeploying the capacity freed by the efficiency gain?
If a leadership team cannot answer these questions before implementation begins, they are not pursuing an AI strategy. They are making a technology bet — and the odds on that bet, as the data consistently demonstrates, are not favorable.
The Vendor Relationship Problem
There is a structural reason why organizations find themselves in this position so frequently, and it deserves candid acknowledgment. The AI vendor market is extraordinarily well-funded, aggressively marketed, and deeply incentivized to accelerate purchasing decisions. Vendors are skilled at demonstrating capability in controlled environments, presenting compelling ROI projections, and creating urgency through competitive framing.
None of this is nefarious. It is simply the nature of a technology market at peak growth phase. But it creates an asymmetry of information and incentive that organizations must consciously counterbalance. The vendor's interest is in a signed contract. The organization's interest is in a measurable business outcome. These are not the same thing, and conflating them is among the most common and costly errors in enterprise technology decision-making.
Independent strategic counsel — advisors without a stake in which platform is selected — provides the counterweight that internal teams often cannot supply on their own. The value of that perspective is not abstract. It is the difference between an AI initiative that generates a line item on a vendor's case study page and one that generates a measurable improvement in your operating results.
Technology Follows Strategy — Not the Other Way Around
The organizations that are generating genuine, documented value from AI share a common characteristic: they approached the technology as an implementation of strategy, not as a substitute for it. They identified specific, high-value problems. They defined success in measurable terms before deployment. They invested in the organizational capabilities required to sustain the change. And they evaluated technology options against strategic criteria rather than feature lists.
This approach is less exciting than announcing an AI transformation initiative at an all-hands meeting. It requires more patience, more analytical rigor, and more willingness to challenge internal assumptions. But it is the only approach that reliably produces the outcomes that justify the investment.
At VW Kumar Consulting, our perspective on AI is neither enthusiastic nor dismissive — it is strategic. We help organizations determine whether AI is the right solution to the problem they have actually identified, design the implementation and change management approach required to generate real returns, and establish the measurement infrastructure to know whether those returns are materializing.
The technology will continue to advance. The organizations that benefit from it will be those that lead with strategy, not with software.
