Gen AI ROI Is Rising, So Why Isn’t Yours?

Enterprise AI has reached the point where “interesting but unproven” is no longer a universal excuse. Deloitte’s own research shows the share of generative AI users reporting significant, measurable ROI rose from 15% in 2025 to 20% in 2026. Real value is now showing up, and the gap between organizations that are capturing it and organizations that are not is getting harder to ignore.

If one in five organizations can already point to significant, measurable ROI, then the problem is no longer whether GenAI can create value, it's why so many enterprises are still failing to convert a very real capability into business impact. At some point, the explanation stops being about the technology and starts being about the organization.

Too many companies are still treating GenAI like an access problem: get the license, send the announcement, offer a few training sessions, then wait for value to emerge. The organizations realizing ROI are doing something different. They are deciding where AI belongs, what work it should change, and what outcomes it is supposed to move.

The conversation needs to shift from adoption to discipline. The companies getting closer to ROI are willing to make AI matter inside actual workflows. They know that productivity gains are not the same as business impact, and they are not confused by a nice-looking demo. Deloitte’s latest report makes that distinction plain: improving productivity and efficiency is where most organizations start, but revenue growth remains much harder to claim. That is the difference between using GenAI and operationalizing it.

The warning sign is easy to spot. If your team is enthusiastic about AI but cannot name the business result it improved, you are probably not behind on technology. You are behind on translation. That is how organizations end up in the Activity Trap, where the tool is busy but the business is not better.

The uncomfortable truth is that the market has moved on, even if some internal programs have not. If ROI is rising elsewhere and your enterprise still cannot point to a meaningful win, the issue is not that GenAI is too early. The issue is that your rollout is too vague.

Enterprise AI is no longer a proof-of-concept exercise. It is an operating choice. If your organization is not realizing value yet, the answer is not to wait longer. It is to get much more specific about where the work should change.

On June 11th, I'm hosting a free webinar called "Why Your AI Rollout Isn't Sticking (And What To Do About It)" and we'll talk about why some teams are already turning GenAI into measurable business value while others are still waiting for the payoff. We'll also talk about what leaders can do when the return is not showing up. If your organization is still in the “we’re using it, but…” phase, this will help you figure out what is actually missing.

Register here.