top of page
Search

The Return Gap is Growing

  • Debbie Braden
  • Jun 29
  • 4 min read

It Was There Before The Rollout Started

Two AI figures standing on opposite sides of a chasm — representing the belief gaps between AI strategy and workforce reality.

80% of CEO’s believe their job is at risk by the end of 2026. (1)


A Harris Poll surveyed 900 executives worldwide and found that CEO’s worry about AI is as strong as the employees in their organizations. If their strategy fails, they’re out.


The pressure for AI integration and enablement is real. What’s less clear is the strategy—what most organizations actually want to achieve.


38% of PE firms expect to spend more than half their total budget on AI this year. Three years ago, most were spending between $15 and $50 million. That number is now pushing past $100 million. (2)


The financial bet to improve profitability and accelerate growth is real. But the results aren’t keeping up with the investment. MIT found that 95% of AI pilots deliver zero measurable P&L impact. Not low return—zero. (3)


The gap between bet and the return isn’t the technology–it’s three belief gaps nobody closed before the AI rollout started.


Gap 1: the wrong mental model


When most leaders think about AI, they think about how they use it personally.


I had one executive reference a TikTok chef’s video for how he wanted to use AI. Honest to God, he did. But it could have just as easily been an AI notetaker, or an assistant that drafts meeting agenda, emails, or helps think through a problem.


The problem is that Enterprise AI is something different. It’s a governed system built on assumptions—built by people often far removed from the actual work being done. It doesn’t learn context from a conversation. It doesn’t know what your frontline knows. It pattern-matches on data it was given, by someone who made decisions about why it mattered based on SOPs and assumptions about the work.


The decision to deploy gets made from the wrong or incomplete mental model. And the distinction between personal AI and enterprise AI expectation rarely gets drawn before the plan is set.


Gap 2: the knowledge that walked out the door


Once the plan is designed, redundant workforce headcount is cut to offset the implementation cost.


Before the cuts, nobody listened to the people doing the work. No conversations about what they actually do—the edge cases, the judgement calls, the informal knowledge that never made it into a process document but kept the operation running.


That intelligence left with the people who held it when their positions were cut. The AI framework gets built on a workforce that no longer exists, trained on assumptions about roles that have already changed.


At the same time, CIOs are discovering the accumulated siloed, redundant and poor data management exposed in the process. Research firm IDC predicts that by 2027 CIOs will face 50% higher AI failure rate and rising costs due to poor-quality data. (4)


Bad data and missing knowledge are two sides of the same problem—garbage in because the SOPs were incomplete to begin with, or because the data was never captured from the people closest to the work.


Gap 3: the compounding cost


The cuts have been made, and the expectation is that the remaining workforce will adopt the new technology—accelerating growth, innovation, and return on investment.


But the employees who watched their colleagues get cut for AI are operating from a new set of assumptions. They may be grateful they still have jobs, but their confidence in AI and their trust in leadership has changed. They are going to do exactly what’s required of them and no more because they learned they could be next.


Manpower Group surveyed nearly 14,000 workers across 19 countries. Regular AI use increased 13% in 2025. Confidence in its utility dropped 18%. (5) Usage went up, belief went down.


Nobody asked what the people believed before the rollout started. Nobody is asking now.

You may be thinking, I don’t need to know what they believe, I need them to do their job. But the lack of alignment shows up as slow adoption, quiet resistance, and an AI investment that doesn’t pay off the way the modeling said it would.


When things don’t go to plan, companies are left rethinking what went wrong. Orgvue found that 55% of businesses admit they made the wrong decision when making employees redundant while bringing AI into the workforce. (6)


The question nobody is asking


The organizations getting real return from AI—Boston Consulting Group puts it at 6%. They aren’t deploying faster, they’re building differently. They found that 70% of AI success is people, process, and change. And the way they define adoption success starts with these gaps—a rich understanding of what employees actually think, feel, and believe. (7)


Which means before the next phase of deployment, before the next round of decisions about where AI goes and who it replaces, there’s a question worth asking.


What do the people inside your organization actually believe about what’s happening? Not in an employee survey or through questions at your town hall. But what they say to each other in the hallway or at home at their dinner tables?


That’s the intelligence most organizations are deploying without. And right now, with board pressure at a peak and runway shortening, it’s the most expensive gap in the room.


Comments

Rated 0 out of 5 stars.
No ratings yet
Add a rating
bottom of page