When AI Hype Meets Enterprise Costs

Key quote:

“AI is a powerful force multiplier, but the judgment, relationships and accountability remain human, and that won’t change.”

Thompson-Reuters is out with the Future of Professionals Report 2026, and it’s confirming some larger trends around the use of in consulting, professional services, tax prep, and related industries. It turns out clients want AI to provide benefits to them, just they’re not seeing it. And similarly, it turns out employees want to know what the strategy is (remember tokenmaxxing?), just… they’re not getting it, either. Which means professionals are using shadow AI, often exposing their organizations to additional hidden risks (Community Bank comes to mind) while trying to meet vague and increasingly expensive objectives.

According to the report, 78 percent of corporate clients now think AI-enabled quality improvements are what they’re paying for, but only 6 percent feel they’re receiving them from providers. One-third of those clients are reconsidering relationships with their providers, estimated at over $1 million in annual spend at risk per firm. That’s not gradual erosion – that’s an H2 2026 revenue risk for law firms and consultancies betting everything on AI differentiation.

Earlier this year I’d covered a Harris Poll that found 71% of CIOs said their AI budgets would be cut or frozen if targets miss their goals by June. 85% said their compensation was tied compensation to measurable AI outcomes. It’s nearly the end of June, and given the price rises in tokens, I can’t imagine July’s going to be fun times for CIOs.

And employees don’t quite know what to do. According to the report, at firms running named AI strategies, 66 percent of professionals say AI “meets or exceeds value expectations”. Where there’s no active strategy, that figure collapses to 22 percent. But having a strategy document isn’t enough. Among those with formal plans, 35 percent say daily work doesn’t match the stated direction because people weren’t trained properly, tools never deployed, or nobody shared understanding of the actual plan.

For instance, Walmart gave employees unlimited AI tokens, usage exploded, and they had to pump the brakes. Accenture said that promotions were tied to AI usage, so they used AI to convert PDFs to markdown, which is something you can just do in your browser, and so they’ve started limiting AI access. Which just pushes workers towards shadow AI. And that’s fun; under Heppner, AI prompts are now discoverable evidence in litigation and antitrust investigations, and shadow AI doesn’t have a magical exclusion.

Unless something changes, clients are going to stop paying firms failing to deliver promised AI value. AI vendors will need to extract maximum revenue before what’s looking like a market correction (disclaimer: I’m still not a financial advisor or an attorney). Employees will keep burning tokens on irrelevant tasks if their performance reviews require it, while leadership will struggle to justify billion-dollar bills. Unless leaders address the gap between what strategies promise and what practitioners actually execute, Q3 and Q4 2026 are going to be difficult at best.