Your biggest AI problem isn’t resistance – it’s false confidence

Matt Shumer’s viral post last week argued the gap between what AI can do and what most people think it can do is “enormous and dangerous.” He’s right. As our CEO Richard Breeden commented, Shumer is “right(ish)” about the speed and scale. What would have taken days and cost thousands is now nearly instant, almost free, and almost as good.

That speed is what makes the risk Shumer pointed out so dangerous. People who tried AI a year or two ago got underwhelming results and concluded it’s not relevant to their role. The same applies to those using free and lower end LLM versions today in comparison with the top models today. Their impression is stuck on a reality that no longer applies to those who have kept up with advances and are taking advantage of the best in class.

But the gap runs both ways. We keep seeing leaders who engage just enough to feel “done”. The question isn’t whether AI works. It’s why so many companies think they’re getting so little from it. In most cases, the use is real and frequent, but basic and optional – so the business doesn’t change.

What adoption looks like when it’s not working

Someone on the team uses an AI assistant 15 times a day and feels like a power user. But they’re operating it as a glorified search engine, so the return feels (and is) marginal. Meanwhile, leadership sees adoption numbers and tells the board AI is embedded. The box gets ticked. What the technology could actually change about the business never gets explored.

At the same time, the way most mid-market firms measure progress makes that false confidence harder to spot. A team researches faster. A report gets drafted in an hour instead of a day. Leadership sees those numbers and assumes the investment is working.

But time saved on individual tasks – real as it is – doesn’t tell you whether AI is changing how your business operates. In most companies we work with at aibl, it isn’t. The tools aren’t connected to the systems where work actually happens. A pilot stalls because nobody can tie it into the CRM.

When the metric is time saved and the tools sit outside real workflows, results flatten. Flat results give everyone permission to stop pushing.

Why nobody fixes it

Leadership sees acceptable metrics and stops pressing. One organisation running an AI upskilling programme for SMEs found something telling. Nearly a quarter of people who signed up and paid never logged in. The non-starters were disproportionately senior – MDs, board directors, heads of department. The people with the most influence over whether AI gets embedded are the ones who feel they already know enough. What they saw six months ago bears almost no resemblance to what’s possible now.

Nobody owns end-to-end change. One advisor working with a 6,000-person company told us she’d been in the role for weeks before hearing the word “agentic” once. The board was pushing for AI strategy and the CEO was supportive. But nobody owned the job of turning pilots into redesigned processes. Teams pick different platforms with no shared standards. Without ownership, conversations stay stuck on basic productivity.

That absence at the top filters down. When leadership isn’t pushing, middle management has no reason to change. Managers default to familiar explanations – leadership hasn’t committed to a platform, it’s not part of my core role, we have security concerns. One transformation leader told us those excuses are usually a signal. The manager hasn’t had the hard conversation with their team about how roles are genuinely changing.

The result is a company that looks like it’s adopted AI but hasn’t changed how it operates.

How to shift AI from cosmetic to operational

Audit where you actually are, not where your board deck says. If adoption means individuals using free tools for tasks they used to do manually, that’s uncoordinated experimentation. Map which teams are using what, whether it’s sanctioned, and what they’re actually doing with it. Most companies we work with at aibl are surprised by what they find.

Get leadership on current models. The tools available today are materially different from six months ago. A leader who hasn’t used the current generation cannot set realistic expectations for what AI should deliver.

Measure business outcomes, not activity. One transformation leader we spoke to spent seven months shifting her organisation from tracking hours saved to quantifiable outcomes – revenue generated, pipeline created, documentation that didn’t exist before. Of 30 AI projects, 23 now have measurable business results attached. We keep it simple: name the process, set a baseline, set a target, review in 30 days.

Close the middle management gap. These are the people who make AI work in practice. Some haven’t engaged at all, others use it daily but haven’t moved past basic prompting. Both groups need hands-on sessions with real workflows, not slide decks about AI strategy. The goal isn’t familiarity with AI – it’s connecting it to the systems their teams actually use.

The question worth asking isn’t whether your team uses AI – it’s whether anything runs differently because of it, not the same processes done slightly faster. If the answer is no, the gap between where you are and where the technology can take you is growing every month. That’s the gap we spend our time on at aibl.