- Most organizations still prioritize AI pilots and point solutions over process transformation and alignment with an enterprise-wide strategy and business goals.
- Returns on AI spend are failing to meet expectations due to several inhibiting factors — and the progression to agentic AI is increasing the pressure.
- Five clear steps can enable organizations to increase and clarify AI ROI and set the scene for enterprise-wide scaling and value growth.
As companies continue to invest heavily in Generative AI (GenAI) and agentic AI, the keys to success are moving beyond pilots to end-to-end workflow transformation, measuring value with fit-for-purpose key performance indicators (KPIs) and applying consistent governance to scale what works.
GenAI and agentic AI continue to attract heavy investment, yet many organizations are discovering that the returns are slower, smaller or more uneven than expected. Leaders are working to build on pilots and proofs of concept, but the value they anticipate is not consistently appearing in the numbers. Instead, they find themselves caught in an AI ROI trap where experimentation accelerates faster than execution, governance and measurement can support.
This article draws on EY research, including two global surveys of technology industry executives conducted with Oxford Economics, to explore why ROI is often fragmentary and hard to prove — and what leaders can do to convert experimentation into sustained value.
About the Research
EY surveys found that about 16% of companies report generating zero ROI on GenAI-enabled Copilot initiatives, and fewer than half (43%) see substantial returns above 50%. This gap between expectation and outcome is starting to challenge GenAI’s ability to deliver meaningful business value.
The issue is broader than capability — it’s timing, readiness, and economics. GenAI’s journey to value is often longer and more complex than expected, with transformation of workflows requiring significantly different investment timelines. While some offer quick wins, achieving sustainable value necessitates rethinking end-to-end processes and platforms.
Many organizations conduct pilots with immature and inconsistent governance that limit scaling beyond proof-of-concept, hindering ROI. The combination of high inference costs — the ongoing, usage-driven expense of running models in production — together with poor model-task alignment and limited change management can prevent even strong prototypes from delivering lasting returns. As AI capabilities advance from copilots to more autonomous agents, these limitations become more consequential — raising both opportunity and risk.
The current state of AI deployment
Organizations take a pragmatic approach to AI, favoring speed, cost and external solutions. But without clear enterprise vision, these choices enforce pilots, create governance gaps and limit ROI.
As our survey results indicate, organizations are taking a wide array of approaches to decisions around AI architecture and delivery, with most choosing to work with external partners. Only 9% are building their own large language models (LLMs), while 18% are co-developing with third-party providers. By contrast, the majority are relying on external models — with 41% using closed models, 27% open and 26% hybrid approaches. Also, customization is largely incremental: 62% use standard external models, while 51% add retrieval-augmented generation (RAG) and 46% fine-tune with proprietary data. The latter two approaches are not mutually exclusive, so some organizations use both.
The high degree of pragmatism in companies’ AI architecture choices flows through to their vendor selection and solution design approaches, as shown in the chart below. On build-versus-buy, the most frequent choice is using a hybrid case-by-case strategy to balance customization with expediency. Out-of-the-box solutions come next, allowing for faster deployment and immediate functionality even if this results in higher upfront costs. Fewer organizations are relying on external providers or building solution-specific AI capabilities in-house. When engaging vendors, 61% of organizations prioritize best-of-breed over best-of-suite approaches.
Read the full report here: How Tech companies can break out of the AI ROI trap
photo source: How Tech companies can break out of the AI ROI trap | EY – Belgium
