The AI Investment Test: Are We Funding Growth or Just Another Cost Program?
The real test of an AI investment is not what the technology can do, but what the business can do differently because of it. Five questions a board should ask before funding growth — or a cost program.
AI has become a boardroom priority. But the real test of an AI investment is not what the technology can do. It is what the business can do differently because of it.
Artificial intelligence has achieved something few technologies ever manage: it has become simultaneously a strategic imperative, a competitive threat, a productivity promise and a line item in almost every executive budget. Boards want to know whether their companies are moving fast enough. CEOs want to demonstrate that they are embracing the opportunity. CFOs want evidence of financial returns. And technology leaders are being asked to deliver capabilities the business has not always determined how to use.
Sound familiar?
We’ve been here before. Cloud transformation, digital transformation and ERP modernization arrived with extraordinary promises. Many delivered substantial value. Others produced more sophisticated infrastructure without fundamentally changing business economics. AI raises the stakes because technology is advancing faster than many organizations can adapt their operating models. In my previous Growth Mandate article, I proposed a straightforward test for technology transformation: does the investment change the company’s ability to acquire, serve, retain, expand or monetize customers?
I believe that same test should now be applied to AI investments, with additional scrutiny.
Are we investing in AI because we have identified a business opportunity, or are we searching for a business justification because AI has become impossible to ignore?
The distinction matters enormously.
The AI Budget Is Not the AI Strategy
One of the most concerning patterns in technology investment is the tendency to confuse financial commitment with strategic clarity. An organization announces a significant AI initiative. It establishes a center of excellence, purchases enterprise licenses, engages consultants, launches pilots and creates a roadmap. These may all be reasonable steps but none answers the most important question: what specific business outcome are we trying to change?
Consider two hypothetical companies investing in AI-enabled customer engagement. The first wants to deploy AI across sales and customer service. Its objectives include improving productivity, reducing manual work and increasing adoption. The second has identified a commercial constraint. Its Customer Success organization cannot consistently identify expansion opportunities or intervene early enough when accounts show renewal risk. That company invests in AI to identify behavioral patterns, surface actionable insights and enable more effective interventions. Both may implement similar technology but only the second has established a clear connection between investment and business outcome. The first is funding a capability. The second is funding a growth thesis. That distinction should shape everything from investment approval to executive accountability.
The Board Should Demand a Growth Thesis, Not an AI Roadmap
Having operated as a CEO, revenue leader and technology executive, I’ve learned that the most important questions often expose assumptions everyone has quietly agreed not to challenge. AI investment is no exception. When a major AI proposal reaches the board, management should articulate its business thesis in plain language. Not: “We will leverage generative AI to improve enterprise productivity.” Instead: “We believe we can improve customer retention by identifying renewal risks earlier. AI will help our teams recognize patterns, prioritize interventions and improve customer engagement.”
Now we have something worth discussing. We can challenge whether the data is reliable, whether teams have the capacity to act and whether the proposed interventions will influence customer behavior. Importantly, we can determine how we’ll know whether the investment worked. This is the difference between approving a technology initiative and making a capital allocation decision. The board’s responsibility is not to become an AI architecture committee — it is to ensure that management connects technology investment to enterprise value.
The Five Questions I Would Put in Front of Every Board
I would apply five questions to every material AI investment.
1. What business constraint are we removing?
Growth is rarely limited by technology alone. It may be constrained by customer acquisition costs, insufficient product differentiation, slow implementation, weak retention, pricing limitations or an inability to scale service delivery economically. AI can potentially address these constraints but unless we identify the constraint first, we risk automating activities that were never particularly valuable. I’ve seen organizations become remarkably efficient at executing processes that should have been redesigned or eliminated. AI makes that risk even greater. Before approving an investment, boards should understand what prevents the business from achieving its objectives and why AI is the appropriate intervention.
2. What will the business be able to do that it cannot do today?
This is where the growth thesis becomes tangible. Will the company serve previously uneconomic customer segments? Introduce more personalized services? Accelerate product development in ways that improve competitiveness? Will sales teams identify opportunities earlier, or will customers receive better outcomes at a lower cost to serve? The answer should describe a new or materially improved business capability, not simply a technical feature. There is an important difference between deploying an AI assistant and enabling a customer-facing organization to operate at a fundamentally different level of effectiveness. The former is an implementation. The latter may be a competitive advantage.
3. Who owns the commercial outcome?
This is where many promising investments lose momentum. The CIO or CTO may own the platform. The data organization may own the models. A transformation office may own the program but who owns the resulting improvement in revenue, retention, customer experience or profitability? If the investment targets sales effectiveness, the commercial leader must own the outcome. If it targets retention, the accountable business leader must translate the capability into measurable results. Technology teams should own reliability, security, integration and delivery but they should not be expected to manufacture commercial outcomes requiring changes in processes, incentives, customer engagement and management behavior. Shared execution is essential. Shared accountability without a clearly identified outcome owner is usually accountability that disappears when results disappoint.
4. What must change beyond the technology?
This is perhaps the most underestimated element of AI investment. Organizations frequently assume that once technology becomes available, people will naturally incorporate it into their work. Experience suggests otherwise. If AI identifies an expansion opportunity, does the account team have a process for acting on it? If automation reduces administrative work, have managers decided how recovered time should be used? If AI improves forecasting, will leaders actually make different resource-allocation decisions? And if AI accelerates product development, can the organization absorb faster releases without compromising quality or customer trust? These are operating-model questions involving incentives, workflows, decision rights, training and organizational structure. They are also where the return on investment is ultimately won or lost.
5. How will we distinguish genuine value from impressive activity?
AI initiatives produce an extraordinary volume of metrics. Pilots launched. Employees trained. Adoption rates. Automated interactions. Hours theoretically saved. These are useful operational indicators. They are not, by themselves, evidence of enterprise value. A company could achieve excellent AI adoption without improving revenue, margins, retention or customer satisfaction.
Boards should insist on two scorecards. The first measures whether technology is working: adoption, accuracy, reliability, risk, cost, and operational performance. The second measures whether the business is improving: customer outcomes, revenue productivity, retention, unit economics and sustainable profitability. Both matter but the second determines whether the investment was strategically justified.
AI Can Be a Cost Program. There Is Nothing Wrong With That.
Not every AI investment needs to produce revenue growth. Some of the most attractive opportunities involve straightforward efficiency improvements. Reducing processing costs, improving service economics, eliminating repetitive work and increasing operating leverage are legitimate strategic objectives. A business delivering equivalent customer value at structurally lower cost may create substantial enterprise value but leadership must be explicit about what it is funding.
If the objective is cost reduction, build the business case around realized savings, implementation costs, operational risk and sustainability. If the objective is growth, establish the commercial mechanism through which AI will generate additional value. If the objective is both, distinguish them rather than allowing ambiguous productivity promises to support optimistic revenue forecasts. There is a difference between saving money and creating capacity to grow — another difference between creating capacity and successfully converting it into growth. Boards should not treat these outcomes as interchangeable.
The Danger of the AI Growth Narrative
The language surrounding AI makes almost every investment sound transformative but transformation is not an economic outcome. Consider an enterprise software company investing in AI-assisted product development. Engineering productivity improves. Teams generate code, documentation and test cases more quickly. Development cycles shorten. That sounds like success but suppose the company continues prioritizing features that do not improve adoption, retention, or competitive differentiation.
It may now be producing the wrong capabilities faster. The same applies to sales.
AI can generate outreach, research prospects, and prepare proposals. But if positioning is weak, product-market fit is poor, or the commercial model is ineffective, increased activity may simply mean more ineffective activity.
AI amplifies the operating model it enters. It does not automatically correct weaknesses already embedded within it.
The board’s investment test must therefore extend beyond technical feasibility. The company must demonstrate that the capability being accelerated is worth accelerating.
Capital Allocation Requires the Courage to Say No
One of executive leadership’s defining responsibilities is deciding what not to fund. This becomes especially difficult when technology attracts extraordinary attention. Every function can identify plausible AI use cases. Every vendor demonstrates compelling capabilities. Every executive can make a case for experimentation. But capital, leadership attention, implementation capacity, and organizational appetite for change remain finite. The danger is not necessarily that companies will invest too much in AI.
It is that they will spread investment across too many initiatives without sufficient concentration to change business performance. I would rather see three well-supported AI investments connected to critical business outcomes than thirty disconnected pilots without a clear path to measurable value.
That requires executive discipline. Boards should resist evaluating progress through portfolio size. The question is not how many initiatives are underway. It is whether the initiatives that matter are changing the business. Leadership should establish conditions under which investments will be scaled, redesigned, or stopped.
Stopping an AI initiative that cannot demonstrate value is not strategic failure. Continuing to fund one because the organization has become politically or emotionally committed to it may be.
From AI Investment to Enterprise Value
I believe organizations capturing the greatest value from AI will not necessarily spend the most or deploy the most sophisticated technology. They will maintain the clearest connection between investment decisions and business outcomes. They will know which constraints they are removing, what capabilities they are creating and which leaders own the results and they will measure whether the promised value materializes. There is one additional issue deserving its own discussion. When AI releases organizational capacity, who decides what happens to it? Does it become cost savings? Fund growth? Improve customer experience? Or disappear into existing workloads? That decision may determine whether an AI investment produces incremental efficiency or a fundamentally stronger business — it is a leadership decision, not a technology decision.
The Growth Mandate
AI does not change the fundamental responsibilities of executive leadership. It makes them more consequential. Boards must allocate capital intelligently. CEOs must establish priorities. Commercial leaders must deliver customer and revenue outcomes. Technology leaders must build reliable capabilities that enable execution. The technology may be revolutionary. The discipline required to turn it into enterprise value is not. Before approving the next AI investment, I would ask one final question:
If we removed the words “artificial intelligence” from this proposal, would the business case still deserve our capital?
If the answer is yes, we may have identified a genuine investment opportunity. If the answer is no, perhaps we are funding enthusiasm rather than strategy. Enthusiasm, however well intentioned, is not a growth thesis.
Nina Simosko writes The Growth Mandate, sapperment’s column on Technology-Led Growth. She is Chief Revenue Officer at CYPHER Learning; former President & CEO of NTT Innovation Institute (NTT i3), Nike global technology leader and SAP Senior Vice President. She writes here in a personal capacity; views are her own. Articles are vendor- and firm-neutral.
Next: read the companion piece, The Capacity Dividend · meet Nina Simosko and The Growth Mandate · or browse the sapperment library.