The current AI conversation, at its most optimistic, treats these tools as a solution to organizational problems. Better communication, better documentation, better decision support — all of it waiting to be unlocked by the right software implementation. That’s mostly wishful thinking, and it’s worth being direct about why.
AI tools are most useful in organizations where the operational foundation is already sound. Where decisions are made by people with clear authority, where information travels reliably, where accountability is real rather than aspirational. In those organizations, AI can genuinely extend capacity. It can automate documentation that currently requires human attention, surface patterns in data that people don’t have time to look for, and help draft communications that would otherwise eat time nobody has.
In organizations where the operational foundation is weak, AI tends to make things faster and more complicated without making them better. Communication tools that move information faster also move misinformation faster. Automation applied to broken processes automates the broken parts. AI that assists decision-making in organizations where decision authority isn’t clear just produces more output that nobody knows what to do with.
I’ve started seeing this play out in engagements where organizations invested heavily in AI tools before addressing the structural problems underneath. The tools added friction instead of removing it. The people using them still didn’t know who was responsible for what. The decisions still weren’t getting made by the right people. The information still wasn’t reaching the people who needed it. The AI just created more surface area for the same problems.
The sequence matters. Fix the operational foundation first. Build clear accountability, reliable communication structures, and decision-making processes that actually function. Then layer tools on top of something solid. In that order, AI is genuinely useful. In the reverse order, it’s an expensive distraction.