I was at a technology offsite recently where the CEO opened with a slide that read: “AI-First Enterprise by 2027.” The boardroom applauded. There was genuine energy.
Two hours later, I was in the canteen talking to a group of senior software engineers and finance managers from the same company. One of them said, quite matter-of-factly, “We’ve been told we have to use the new AI assistant for all our report generation and code checks now. Nobody asked us if it actually makes sense. It adds three extra verification steps to my day. So, I just open it, run a dummy prompt to keep the metrics team happy, and then do the work the way I’ve always done it.”
Same company. Same day. Two completely different realities.
This is the AI trust gap − and it is the primary reason why BCG’s 2025 research found that only 5% of companies are actually generating value from AI at scale.
The technology works. The demos are flawless. The deployment dashboard shows “90% adoption.” But the actual business output isn’t shifting. Why?
Because the metric we are measuring − “adoption” − is a proxy. And under pressure, proxies lie.
Frontline staff aren’t resisting AI because they are afraid of the future. They are resisting Tool Fatigue. They are rejecting half-baked corporate AI wrappers that add administrative friction to their daily tasks. When you layer an automated tool onto a broken, legacy process without involving the people who actually run it, you don’t get efficiency. You get silent resistance. Your people perform digital theatre to satisfy executive dashboards, while the real work goes on in the background, completely untouched.
To close this gap, enterprise leaders must stop measuring activity and start measuring Workflow Integrity.
The organisations successfully scaling AI do three things differently:
They ban “Digital Theatre” metrics: They stop measuring “seat logins” or “monthly active users.” Instead, they measure end-to-end outcome metrics: change in cycle time, drop in error rates and task throughput.
They redesign the workflow, not just the tool: They do not hand down tools from on high. They put the deployment team in the room with the operators to co-design the new process, stripping out legacy steps rather than adding new ones.
They leverage “AI Pilots”: They identify the natural builders within the teams − not managers appointed from above − and give them the autonomy to adapt the tools to actual frontline realities.
At Experis, this gap between license deployment and operational reality is where we live.
We don’t believe in “body-shopping” or dropping off licenses and hoping for the best. Our Project Services teams tie our delivery milestones directly to the SLA of operational adoption. Supported by our cross-border delivery hubs, we help you redesign your processes, deploy the right specialists and ensure your AI investments translate into measurable gross margin, not just vanity slides.
Here is the question I would put to any technology leader celebrating high AI adoption numbers: Are your people actually doing better work, or have they just learned how to click the buttons you’re measuring?
Next in the series: Post 2 − AI is eating your junior talent pipeline. And nobody’s noticed.
References:
BCG: To Unlock the Full Value of AI, Invest in Your People (2025) (https://www.bcg.com/publications/2025/to-unlock-the-full-value-of-ai-invest-in-your-people)
BCG: AI at Work — From Potential to Profit (2025) (https://www.bcg.com/capabilities/artificial-intelligence/overview)
