Most enterprise AI ROI is lagging for a simple reason. Companies are buying capability faster than they are redesigning workflows.
The survey data is consistent. McKinsey (November 2025): only 39% of organizations report any EBIT impact from AI, with most under 5% of EBIT. BCG (September 2025): 5% of companies are genuinely future-built, while 60% remain laggards.
The real question is whether AI usage is translating into measurable outcomes: faster cycle times, lower servicing costs, tighter staffing ratios, higher revenue per employee, and better margins.
In financial services, the gap is visible. Analysts can use AI to draft memos faster, summarize datarooms, and pull comparable structures. That is real productivity.
But the cycle time does not change. The approval workflow does not change. The committee process does not change.
Speed at the task level disappears into the same process.
User-level adoption can create enthusiasm. Workflow-level adoption is what creates durable cash flow impact.
The firms that matter over the next few years will not be the ones with the most pilots. They will be the ones that convert AI from an employee tool into a redesigned operating model, and then show the result in the numbers.
That is the area worth underwriting.