EY: Escape the AI ROI Trap with Strategic Actions

The AI ROI Trap: Why Tech Firms Struggle

Organizations are heavily investing in Generative AI (GenAI) and agentic AI, yet many are failing to realize expected returns. This puts them in an ‘AI ROI trap’, where experimentation often outpaces execution, governance, and measurement capabilities.

EY’s research, conducted with Oxford Economics, involved global surveys of technology executives to understand these challenges.

Key Challenges and Lagging Returns

• Approximately 16% of companies report zero ROI on GenAI-enabled Copilot initiatives.
• Fewer than half (43%) achieve substantial returns above 50%.
• This gap stems from prioritizing small-scale pilots and point solutions over comprehensive, end-to-end business transformation.
• Many organizations conduct pilots with immature and inconsistent governance, limiting their ability to scale beyond proof-of-concept.
• High inference costs, poor model-task alignment, and limited change management also prevent prototypes from delivering lasting returns.

Current AI deployment strategies often prioritize speed, cost, and external solutions. Most organizations rely on external models, with customization largely incremental rather than building their own Large Language Models (LLMs).

ROI is also lagging due to insufficiently defined Key Performance Indicators (KPIs) and inadequate governance frameworks. A significant 61% of respondents believe AI implementation creates more value than they can accurately quantify.

EY’s Five Actions to Escape the Trap

To escape the AI ROI trap, organizations can take five key actions:

1. Scale AI through processes, not tools: Redesign end-to-end workflows to capture value at the process level, scaling only pilots with tangible ROI and strategic fit.
2. Define value, measure it, and embrace iterative review: Shift to fit-for-purpose KPIs tied to business outcomes like growth, resilience, and cost-to-serve.
3. Strengthen governance: Progress to consistent, enterprise-level decision rights integrated with deployment, operations, and measurement.
4. Invest in AI-ready data with integrated risk management: Ensure robust data lineage, stewardship, and quality, making data readiness a gating criterion for scaling AI.
5. Manage AI as a strategic portfolio: Balance foundational investments with experimental projects, adopting a dual mindset for predictability and transformation.

Outlook: Treating AI as a Strategic Investment

Ultimately, realizing AI’s transformative potential requires disciplined capital allocation, strong foundations, and portfolio rigor. Organizations must treat AI as a strategic investment anchored in readiness, rigor, and measurable results.