AI Bites: When AI becomes your operational x-ray

 

ITG’s Intelligent Transformation Director, Ben Gallagher, on why businesses need to get their operations in order if they want AI to create meaningful impact. 

AI projects rarely fail because the model isn’t good enough. They fail because, for the first time in decades, companies are confronted with the reality of how work actually gets done — not how it appears on a process map, not how leaders believe it happens, and definitely not how the documentation describes it.

Before AI, organizations could survive on approximations. “Roughly how it works” was good enough. “Just ask Phil, he knows” was a perfectly acceptable workflow. Humans are exceptional at bridging gaps, interpreting ambiguity and compensating for systems that were never designed to handle real‑world complexity. Almost every business relies on these invisible acts of ingenuity to function.

AI doesn’t. And that’s where the disruption begins.

When you try to automate even a simple task, all the hidden variability surfaces. Tasks that look identical on paper turn out to be completely different in practice. “Minor” exceptions turn out to be the norm. Fields in the same system mean different things in different teams. Processes contain “optional” steps that everyone quietly knows are absolutely critical. Rules exist, but no one can articulate them because they’ve lived for years inside people’s heads.

You don’t notice this when the work is human‑powered. People smooth over gaps without even realizing they’re doing it. They infer meaning from context. They compensate for incomplete inputs. They bring a level of judgement and tacit knowledge that has never been captured, formalized or standardized.

But ask AI to do the same thing and suddenly the chaos is visible. Painfully visible.

The truth is that AI doesn’t create disorder — it exposes it. It’s less a tool and more an operational X‑ray. And X‑rays are uncomfortable if you’ve spent years pretending the bone wasn’t broken.

This is why most AI programs quietly evolve into process redesign projects. The moment teams begin mapping what AI is supposed to do, they discover that no one actually agrees on the “right” way the work should flow. Leadership’s version, the process documentation’s version and the real‑world version rarely match. The organization is forced to confront the mess that humans have heroically navigated for years.

That confrontation is healthy. Necessary. Transformational.

Because when you fix the underlying complexity — when you standardize, clarify, document, streamline and align — AI doesn’t just work better. The whole organization works better. Decision‑making speeds up. Errors drop. Teams experience fewer handoff headaches. Technology stops fighting the business, and the business stops relying on heroic individuals to hold everything together.

The companies seeing the biggest return on AI aren’t the ones with the most advanced models. They’re the ones willing to be honest about their own reality. They treat AI as a catalyst for operational clarity, not a magic wand.

AI will reshape the future of work. But first, it reveals the truth of the work we already do. And sometimes, the hardest part of transformation is simply looking at the X‑ray — and deciding what needs to be healed.