Corporate leaders are confronting a sobering reality: most AI investments are not paying off. McKinsey's 2025 Global Survey found that 88% of organizations use AI in at least one business function, but only 39% report any impact on EBIT, and even among those, the impact is typically less than 5%. BCG's analysis reveals that 60% of companies investing in AI generate no material value, and only 5% create substantial value at scale. Deloitte's survey of nearly 2,000 executives finds that satisfactory ROI on a typical AI use case takes two to four years-much longer than the seven-to-twelve-month payback typically expected for technology investments.
These numbers signal a fundamental mismatch between how companies invest and what AI actually delivers. The problem, according to recent analysis, is not the technology itself but the framework used to evaluate it. Traditional ROI metrics fail to capture the distinct value paths that different AI investments create.
The five categories of AI investment
AI investments fall into five distinct types, each with its own logic for value creation. Two are tactical, designed to maintain competitive position. Three are strategic, built to create durable advantage. None can be measured against conventional ROI in the same way.
The first tactical category is process automation. This involves using AI to replace or accelerate repetitive, rule-based tasks. Value here comes from cost reduction and efficiency gains. The second is cognitive augmentation, where AI assists human decision-making without replacing it. These systems surface insights, flag anomalies, or recommend actions while leaving the final call to people.
The three strategic categories start with product enhancement. AI gets embedded into existing offerings to make them smarter or more personalized. Then there is new offering creation, where AI enables products or services that were previously impossible. The final, deepest layer is business model transformation, where AI reshapes how the organization creates, delivers, and captures value altogether.
Why traditional ROI fails
Applying standard capital-budgeting logic to AI investments breaks down for clear reasons. Tactical investments in automation and augmentation can show near-term efficiency gains, but their real value often lies in freeing up organizational capacity that traditional accounting never captures. A process made 30% faster might not change EBIT directly-it might simply allow the same team to handle more complexity without burning out.
Strategic investments are even harder to measure. Product enhancements may increase customer retention or willingness to pay, but those effects ripple through financial statements over years, not quarters. New offerings and business model shifts require patience and a willingness to fund exploration that standard hurdle rates actively discourage. For those pursuing AI for Executives & Strategy, this demands a different mental model for resource allocation.
The value capture challenge
Even when AI creates real economic value, companies struggle to capture it. The BCG finding that only 5% of firms generate substantial value at scale is not primarily a technology failure. It reflects organizational gaps: data infrastructure that cannot support production-grade AI, talent models that treat AI as an IT project rather than a business capability, and governance structures that slow deployment to a crawl.
Deloitte's two-to-four-year ROI timeline is not a warning sign-it is a realistic benchmark for investments that require organizational learning. Companies that pull funding after 12 months because they see no EBIT movement are essentially abandoning the effort before the value curve bends upward. The firms that succeed treat AI investments less like software purchases and more like R&D programs with staged funding, clear decision points, and executive patience.
Why this matters for executives and strategy leaders
Executives evaluating AI investments need to discard the single-ROI-number mindset. The five categories-automation, augmentation, product enhancement, new offerings, and business model transformation-each demand different success metrics, different time horizons, and different governance. A process automation project that does not pay back in 18 months is a failure. A business model transformation that does not show returns for five years may be on track. The skill is not in calculating a better number. It is in knowing which category you are funding and holding yourself to the standard that actually fits. For leaders in AI for Finance and investment management roles, this means building evaluation frameworks that distinguish between these types rather than treating all AI spend as interchangeable line items.
Your membership also unlocks:
