Set realistic expectations before investing in AI

Prolego's view is that more than 95% of major enterprise AI investments fall short of their original expectations. We suspect the proportion is closer to 99%. This is an opinion grounded in our experience working with enterprises, rather than a quantitatively established industry rate.

The practical implication is to enter an AI investment with a clear understanding of the risk. Substantial spending, visible activity and a deployed application can all coexist with disappointing business impact. Leaders should recognize that possibility before committing resources and structure the work to learn quickly.

Compare the impact with the original investment case

The relevant comparison starts when an organization commits money, operating resources or engineering time. What change justified that commitment? What did the sponsors expect the investment to accomplish? How does the delivered impact compare with those expectations?

A system can reach production, attract users and improve a workflow while delivering much less than the business case anticipated. Our view concerns that gap. It does not mean that 95% of projects fail technically, never deploy or produce no value.

Keeping the original expectations visible matters. Otherwise, a project can gradually redefine success around what it managed to deliver, leaving the larger investment rationale unexamined.

AI activity is outpacing business transformation

Across enterprise environments, we see organizations buying agentic frameworks, hiring full-time teams, deploying workflows, running vendor pilots and expanding access to tools such as Claude Code and Copilot. The activity is substantial. In our experience, the resulting impact repeatedly falls short of what the investment was expected to achieve.

Each initiative may have a plausible purpose. Together, they can produce fragmented capabilities without materially changing how the company operates or makes money. A collection of deployed solutions does not automatically compound into a more capable business.

That is why our starting point is re-engineering business economics. The business ambition should guide the technology, team and first investment.

Industry research provides context

Published surveys also distinguish widespread adoption from broader impact. McKinsey's November 2025 survey reported regular AI use in at least one business function among 88% of respondents, while 39% reported an impact on enterprise earnings before interest and taxes. These are survey-reported outcomes. McKinsey's 2025 survey.

In September 2026, Gartner reported that 22% of surveyed organizations had scaled AI across multiple business units or adopted an AI-first approach. Its high-performing group also reported positive returns in 81% of initiatives. Gartner's 2026 survey.

Adoption, positive returns, enterprise impact and meeting the expectations used to justify an investment are different measures. The research helps illuminate those distinctions; it does not establish our opinion's percentage.

Think big, start small and reduce risk early

Falling short should be treated as a normal investment risk, not an exceptional outcome that only poorly managed companies face. The response is to examine why an opportunity should work and make the first commitment small enough to learn and adapt.

This does not require certainty before beginning. AI is advancing too quickly for prolonged planning, and waiting carries its own risks. It requires a substantial business ambition, a practical first investment and a method for exposing the assumptions that matter.

Performance-Driven Development brings reliability risks forward. The first-month performance report helps stakeholders understand expected behavior, potential impact and remaining uncertainty so they can decide whether to continue, adjust or redirect the work. Early risk mitigation explains that approach.

Once a useful system is in operation, production feedback and rapid improvement develop the capability further. Each investment should create value while contributing to the larger transformation.

Discuss the opportunity before committing resources

We are happy to offer initial advice before a company funds a pilot, hires a team, buys a framework or selects a vendor. A company name and a sentence or two about the idea are enough to begin.

Email kevin@prolego.com or russ@prolego.com. We can research the larger opportunity, likely headwinds and a practical starting direction, then discuss the findings with you. Our scoping approach describes the process.