Companies Using AI to Cut Costs Are Missing the Biggest Value-Creation Bet of the Last Decade
There is a gap between what executives say they believe about artificial intelligence and what their organizations actually do with it. It is not a knowledge gap. It is a strategic attention gap, and it carries a cost that few boards of directors have quantified with genuine honesty.
At a recent roundtable with executives from the wealth management sector, the authors of a Wharton paper posed a direct question: if in three years we were to compare two similar firms — one that made good use of AI and one that did not — how much more valuable would the first one be? The average answer was 2.35 times, equivalent to a 135% increase in firm value. A number that the participants themselves considered reasonable. The problem arose immediately afterward, when they were asked where they were actually investing in AI. The answer was nearly unanimous: in efficiency. Several admitted they had never seriously connected AI to revenue growth.
That is not a vision problem. It is a decision architecture problem.
When the Efficiency Ceiling Becomes a Strategic Ceiling
The case for using AI to reduce costs has empirical support. A large-scale randomized trial at a software company found that a generative AI-based customer service tool increased agent productivity by more than 10%. A separate study involving nearly 5,000 developers showed gains exceeding 25%. In wealth management, AI can compress weeks of client onboarding into days, and assist advisors in meeting preparation and follow-up. Those are real results.
But there is an arithmetic that efficiency models cannot overcome. Under generous assumptions, if 50% of a firm's cost base is susceptible to AI-driven improvements, and AI reduces those costs by an average of 10%, the total impact on expenses is approximately 5%. Applied to a representative wealth management firm, that produces a value increase of around 10%. That is not negligible. But it is a world away from the 135% that those same executives considered achievable.
The reason is structural, not circumstantial. Costs have a lower bound: zero. Revenue has no upper bound. And capital markets do not value companies primarily on what they earn today, but on what they are expected to earn in the future. The premium that investors assign to expectations of sustained growth is disproportionately large compared to the premium they assign to expense optimization. A wealth management firm that grows organically at 5% per year is worth approximately 50% more than an otherwise identical firm growing at 3%. One that grows at 7% is worth 122% more. Those numbers do not emerge from optimistic projections: they are the direct consequence of how markets calculate earnings multiples when sustained growth is on the horizon.
What this implies is that an increase of just two percentage points in the organic growth rate — something modest for historically high-performing firms — can increase firm value by 50% before earnings themselves have even grown. An increase of four percentage points can double that value. Against those magnitudes, savings in operating costs become a second-order argument.
The Experiment That Demonstrates the Mechanics of Growth
To make concrete what might so far sound abstract, the researchers worked with wealth management firms on a specific application: direct marketing campaigns on LinkedIn, targeting senior executives and SME owners. The approach was unconventional.
They used what they called "virtual scientists": AI systems instructed to generate dozens of alternative ad concepts and then simulate the target audience's response in order to identify, before launch, which ones would perform best. The projected increase in click-through rates for the winning ads ranged between 2.7 and 3.5 times. When those ads were deployed in the field, the average increase was 3.2 times.
The relevant question is not whether that number is impressive. It is what it does to firm value. Consider a company with a base organic growth rate of 3%, distributed more or less equally across three channels: advisor networks, purchasing leads from custodians, and direct marketing. If the direct marketing channel accounts for one percentage point of that rate and AI triples it, that channel moves to contributing three percentage points. The total organic growth rate rises from 3% to 5%, and firm value increases by approximately 50% from that single move alone.
Now suppose that, encouraged by those results, the firm redirects the budget it previously devoted to lead purchasing — an expensive source with diminishing returns — toward the direct marketing channel, which has just demonstrated real traction. The organic growth rate jumps to 7%. The firm's value more than doubles relative to the baseline.
This is not a laboratory experiment. It is a demonstration of the mechanism by which the allocation of AI resources determines whether a company captures the 10% in value that efficiency promises, or the 100%+ that growth promises.
The Asymmetry That Most Boards of Directors Are Not Seeing
There is something more disturbing than the fact that companies are underinvesting in AI for growth. It is that competitive dynamics are going to make that gap harder to close over time.
The marketing gains that AI produces today — that 3.2 times in click-through rates — are going to compress as more firms adopt similar tools. The window for capturing valuation multiples from those results is finite. What does not compress at the same speed are the growth levers that depend on relational depth: expanding the share of wallet within existing client relationships, improving the quality of financial advice, shortening sales cycles through better alignment between advisors and client profiles. Those levers are harder to imitate because they require the accumulation of context, trust, and proprietary data.
The firms that first build a foundation of sustained organic growth also enjoy a secondary advantage that few strategic analysis models capture clearly: higher valuation multiples become acquisition currency. A company with a high multiple can acquire competitors with lower multiples with less dilution for its own shareholders. Efficiency does not generate that effect. Sustained growth does.
The argument extends well beyond wealth management. Any sector in which investors value sustained organic growth — from legal services to healthcare, from education to software platforms — faces the same asymmetry: the multiplier effect of growth on valuation far exceeds the impact of cost reduction. The firms that recognize this first do not merely grow faster: they position themselves to define the competitive structure of their sector for the years ahead.
Dependency on the Efficiency Agenda and Silent Structural Fragility
There is a dimension that financial analysis does not fully capture, and which from an organizational perspective matters just as much as the numbers. Organizations that orient their AI agenda primarily toward efficiency are not being conservative. They are building a structural dependency on a type of return that has a ceiling, at a moment when the market is massively rewarding a different type of return that does not.
This creates fragility of a specific kind: not the visible fragility of an indebted company or one with negative margins, but the fragility of a system that functions well within its own parameters and therefore feels no urgency to change them. Costs fall, processes improve, reports show progress. But the organic growth rate does not move, and neither does the valuation multiple.
The trap does not lie in team incompetence or in a lack of technical talent. It lies in the fact that the efficiency program has clear metrics, short feedback cycles, and well-defined internal stakeholders. The AI-driven growth program requires field experimentation, tolerance for results that do not confirm initial hypotheses, and a willingness to redistribute budget from established channels toward capabilities that are still being proven. For many organizations, that redistribution does not run into the technology. It runs into governance, area-level incentives, and the speed at which committees approve experiments that do not fit existing budget categories.
The authors of the Wharton paper call this absorptive capacity: the degree to which an organization's people, governance processes, and workflows can incorporate and act on new technology. For many firms, the first real obstacle to converting AI into growth is not building better tools. It is removing the internal bottlenecks that prevent existing tools from being used effectively at scale.
The most structurally mature organizations are not necessarily those with the most sophisticated technical teams. They are those that have built the institutional capacity to take a piece of field evidence — such as the LinkedIn experiment — and convert it into a resource reallocation decision before the window of competitive advantage closes. That capacity is not installed through a digital transformation project. It is built through repeated decisions about how strategic attention is allocated, what is measured as success, and who has the authority to redirect budget when the data justifies it.
Companies that today are using AI primarily to reduce costs are not making a wrong decision in any absolute sense. They are making the decision that their governance structures, incentive systems, and reporting cycles make easiest to take. The problem is that this ease carries a price that does not appear on any current income statement, but that will appear in comparative valuation multiples three years from now.
