Why Is AI Investment Still Failing to Reach the P&L?

AI technology is increasingly capable. Enterprise execution is the constraint.

McKinsey & Company argues that leading organizations are redesigning how work is performed and how decisions are made, rather than placing AI inside unchanged processes. Gartner similarly finds that finance functions continue to direct most AI spending toward individual productivity and process-improvement use cases rather than initiatives capable of materially changing business outcomes.

The implication is clear: isolated tools may accelerate tasks, but material enterprise value depends on redesigning workflows and the operating systems around them to improve cost, throughput, decision speed or growth outcomes.

Within finance, AI spending remains heavily weighted toward individual productivity and process improvement rather than enterprise-level value creation. According to Gartner, 84% of finance AI spending supports individual productivity and process-improvement use cases, while only 16% supports initiatives capable of materially changing business outcomes.

KPMG adds an important economic control: organizations with full visibility into AI operating costs are five times more likely to report established ROI, at 15% versus 3%.

Enterprise AI value realization therefore depends on cost transparency, connected systems and clear accountability — not model capability alone.

Workforce productivity does not reach the P&L automatically. PwC finds that financial-services firms often observe adoption and productivity before they can identify revenue or cost effects.

Management must decide what happens to released capacity. Released capacity can be redirected toward training, higher-value work, stronger customer relationships or growing areas of the business. Without a deliberate strategy, released time — and the value of that time — may simply leak out.

Leaders must therefore integrate AI investment with workforce planning, organizational and workflow redesign, skills development and financial measurement.

In PwC's 2026 Financial Services Workforce AI Survey of 1,004 US executives, 77% said most AI investments were not delivering measurable ROI. The report recommends tracking adoption and productivity as leading indicators before expecting lagging financial results.

Skills are also becoming more economically important: separate PwC research found that roles requiring AI skills now carry an average wage premium of 62%, while postings for those roles grew 69% against 9% for the market as a whole.

This creates a dual requirement for executives: build and fund the skills needed to scale AI while applying financial rigor to workforce and AI investment decisions.

The most durable AI advantage is unlikely to come from generic productivity tools that competitors can also acquire. It will come from changing how the institution creates, delivers and captures value.

McKinsey & Company argues that organizations must use broadly available technology to build capabilities competitors cannot easily reproduce. PwC reinforces this view by finding that AI leaders focus on growth as well as efficiency.

For banks and insurers, this can include reducing customer-journey friction, scaling core workflows, launching new products more efficiently and pursuing new sources of growth.

PwC found that the top 20% of 1,217 surveyed companies captured 74% of AI-driven returns. These organizations generated AI-related financial performance 7.2 times higher than other respondents.

The concentration suggests that value is not distributed evenly among adopters. It accrues to organizations that coordinate strategy, data, investment, workforce, governance and innovation.

The evidence also supports a strategic distinction: efficiency can reduce costs, while redesigned workflows, stronger customer experiences and business-model reinvention can unlock growth and higher-value outcomes.

Executives need stronger financial discipline, but not every AI initiative should be judged identically. Gartner recommends treating AI as a portfolio of different bets.

Productivity and automation use cases can be assessed through cost, efficiency and productivity outcomes. Growth-oriented use cases should be evaluated against the revenue and customer outcomes they are designed to create. Transformational investments require different targets, timelines and risk expectations, with weak bets stopped early and stronger ones scaled.

Forrester reinforces this approach by showing that conventional measurement often fails to capture AI's different financial, operational and strategic value mechanisms while still requiring defensible executive evidence.

The value gap is visible across multiple surveys. PwC found that 30% of CEOs reported AI-related revenue gains and 26% reported lower costs, while 56% had realized neither outcome and only 12% had realized both.

KPMG reported that many leaders see meaningful business value from AI, while only 7% can point to established ROI — and 24% are already facing investor pressure to demonstrate it.

These results are not directly comparable because the samples and definitions differ. Together, the surveys show that reported business value and measurable financial returns are distinct — and should not be treated as equivalent.

Financial institutions should measure AI through realized operational and financial outcomes — not deployment volume. In banking and insurance, AI value should be demonstrated through measurable improvements in the business workflows where AI is deployed.

In finance operations, leaders should assess whether AI improves close and forecasting processes, execution speed and decision quality.

PwC and Gartner both emphasize that deployment only matters when it produces measurable operational or enterprise value.

Financial-services organizations are investing heavily, but evidence remains uneven. PwC found that 80% of financial-services respondents are scaling AI enterprise-wide or have fully embedded it across business units — the highest of any industry surveyed — while its AI Agent Survey found customer service, sales and marketing, and IT and cybersecurity among the leading application areas.

Gartner found that nearly 60% of CFOs planned to increase finance AI investment by at least 10%, while efficiency remained a dominant objective — staff productivity is a top-three priority for 88% of them.

The sector opportunity is substantial, but productivity gains should be connected to measurable process, financial and strategic outcomes, alongside financial-services risk and control considerations.