October 22, 2026 • 5 Min Read
AI can make thousands of activities faster without materially changing the financial performance of an enterprise.
In June 2026, Gartner reported that 84% of finance organizations had implemented or were planning to implement AI, yet only 7% reported a high or very high impact. Gartner’s July 2026 findings pointed to a related imbalance: 45% of finance leaders said their AI investments leaned toward productivity, compared with 20% toward decision quality, even as boards sought broader enterprise value.
Part of the problem is where organizations look for returns. Making existing activities faster can deliver useful efficiencies, but the greater economic opportunities may lie in better commercial decisions, lower capital requirements or entirely new sources of revenue.
There is also a more fundamental constraint: the organization’s ability to turn AI capabilities into business results.
RGP’s September 2026 survey of 400 CFOs and equivalent finance leaders points to an execution ceiling. Nearly half (49%) say their transformation agenda exceeds their organization’s capacity to execute. Within that group, just 24% report positive AI ROI, compared with 79% where capacity matches or exceeds transformation demands.
For CFOs, these findings raise two connected questions: Where can AI materially change the economics of the business, and does the organization have the capacity to capture that value?
The Capacity Ceiling on AI Returns
The distinction between creating value and capturing it is becoming central to the economics of AI.
The capacity constraint is not necessarily a shortage of employee hours. In RGP’s research, 63% of CFOs say AI has already reduced their finance organization’s total workload, and 47% plan to redeploy employees to higher-value strategic work. Yet the enterprise may still be constrained by systems integration, specialist skills, fragmented accountability or slow decision-making.
Consider demand forecasting. AI might substantially improve forecast accuracy, but the financial benefit depends on whether purchasing, production and inventory decisions change in response. Better predictions alone will not reduce working capital.
The same is true of customer intelligence, financial planning or risk management. AI can produce better information, but capturing its economic value may require different processes, incentives, decision rights and ways of working across functions.
This changes what belongs in an AI investment case. Alongside the expected performance improvement, CFOs need to understand what the organization must do differently to realize it, who will be responsible and whether the necessary changes are achievable.
Different Opportunities Have Different Economics
Not all AI investments create value in the same way, and they should not be evaluated against identical expectations.
Applications that automate established activities generally have a relatively direct relationship between costs and benefits. The economics of invoice processing, reconciliation or routine reporting are often well understood, making returns comparatively straightforward to estimate.
Investments that improve decisions are different. Their financial benefits may come from reducing uncertainty, improving resource allocation or making the enterprise more responsive. The returns can be substantially larger, but the relationship between the technology and the financial outcome is less direct.
AI investments that enable new products, services or business models introduce another set of considerations. Their value may lie in opening new markets, creating revenue streams or making offerings commercially viable that previously were not.
These distinctions have important implications for capital allocation.
An efficiency initiative might reasonably be expected to produce measurable savings within a short period. An investment in enterprise-wide planning could require a longer horizon and coordinated changes across several functions. A new AI-enabled service should be evaluated much like any other growth investment, with assumptions about market demand, adoption, profitability and competitive differentiation.
Applying identical payback requirements to all three can systematically favor smaller, more predictable projects over opportunities with greater long-term economic potential.
The challenge for CFOs is to maintain investment discipline while applying measures, horizons and milestones appropriate to the different sources of value.
The Full Economics of AI Investment
The technology itself may represent only a fraction of what it costs to capture the benefits of AI.
Data quality, systems integration, security, governance, model oversight and organizational change all influence the economics. Some costs may be attributable to individual projects, while others support multiple investments across the enterprise.
This creates an important portfolio consideration. Several promising applications may depend on the same data architecture or integration capabilities. Evaluated individually, each might appear too expensive. Considered together, they may justify shared infrastructure that supports a much broader set of returns.
The reverse is also possible. Individual projects may appear attractive until their cumulative infrastructure, governance and operating costs are considered.
In RGP’s research, integrating AI into existing systems is the most frequently cited single barrier to AI value, at 23%, followed by scaling successful pilots at 14%.
The investment question therefore extends beyond the economics of individual applications. CFOs need to understand where capabilities can be reused, where investments reinforce one another and whether connecting applications across functions could create value that none could deliver independently.
Manage the Portfolio Around Realized Value
Many organizations have built their AI portfolios through experimentation. The next stage requires greater discipline in how investments are prioritized, sequenced and managed.
When execution capacity is limited, a portfolio of individually attractive projects may still be impossible to deliver. Initiatives compete for the same specialist talent, leadership attention, integration resources and organizational capacity for change.
The highest-return decision may therefore be to defer an additional AI application and instead remove a constraint that is limiting several existing investments. This makes sequencing as important as project selection.
Each significant investment should have a clear economic hypothesis: what aspect of business performance will improve, by how much and through which operational changes? There should also be credible baselines, measurable milestones and explicit responsibility for realizing the benefit.
Governance matters. RGP’s research finds that CFOs at organizations executing major initiatives through enterprise-wide governance with a single point of accountability are more likely to report significant positive AI ROI than those in siloed organizations: 22% versus 8%.
As evidence accumulates, some initiatives will warrant acceleration. Others may need redesign because the technology works but the business has not adapted to it. Still others should be stopped when their economics no longer support continued investment.
The urgency is clear. Just 10% of CFOs in RGP’s research report significant positive AI ROI today, while 54% expect it within 24 months.
Meeting those expectations will require more than additional investment. It will require an operating plan that addresses the constraints preventing existing and future investments from delivering their full economic value.
A Broader Role for the CFO
AI is expanding the CFO’s role in shaping enterprise performance, but there is a potential mismatch between responsibility and authority.
In RGP’s research, 64% of CFOs report high accountability for AI adoption, while only 18% report high control over its execution.
That gap matters because the financial returns from AI increasingly depend on operating decisions that extend well beyond finance. Capturing them requires agreement among the CEO, CFO and functional leaders about priorities, investment tradeoffs and ownership of the changes needed to achieve results.
The opportunity for CFOs is to bring an enterprise-wide economic perspective to those decisions, determining not only which investments deserve capital, but which combinations of investment and organizational change can produce the greatest financial return.
Ultimately, the question is not how much work AI can eliminate. It is how fundamentally AI can change what the enterprise is capable of achieving—and whether the organization is prepared to realize that potential.