From Insight to Instinct, Turning Patterns into Tools

Dashboards get a lot of credit they haven't earned. I say that as someone who has built more of them than I'd like to admit, and who still believes they have a place. But I've watched too many organisations treat a dashboard as the finish line, when really it's only ever a waiting room.

Here's the pattern I've seen more times than I can count. Somebody notices something isn't quite right. Service is a little slower on Friday afternoons, a particular menu item keeps coming back, a certain shift always runs short on cover. A dashboard gets built to track it. Everyone's pleased. The dashboard gets checked for a while. And then, gradually, it stops being checked, because knowing about a problem every week isn't the same as solving it, and eventually the human brain stops paying attention to information that never asks anything of it.

That's the moment worth watching for. Not when a pattern first appears, but when it starts repeating. A single occurrence is just Tuesday having a bad day. A pattern that shows up three weeks running, across more than one site, isn't an anomaly anymore. It's a signal. And signals, in my experience, are asking to become something more useful than a chart. The habit I try to build into every team I work with is a simple discipline of noticing repetition and treating it as a prompt rather than as background noise. If the same thing keeps showing up on a dashboard, the dashboard has already told you everything it's going to tell you. What it can't do is act. That's not a criticism of dashboards. It's just an honest description of what they are, which is a visibility layer, not a decision-maker.

So the question I ask, once a pattern repeats often enough to earn my attention, isn't "how do we display this more clearly." It's "what would it mean for this to tap someone on the shoulder instead of waiting for them to look?" That's a different kind of build entirely. It's smaller, usually. It's more specific. And it tends to land with people in a way that a chart never quite manages, because it arrives at the moment they need it rather than the moment they remember to go looking for it.

This is the difference between telling somebody something and helping them do something. A dashboard tells. A prompt helps. The gap between those two things is, more often than not, where the real value of any internal tool actually lives.

There's an order to how much of this people will accept, by the way, and skipping steps in it costs you dearly. It matters enough that I've given it a whole chapter of its own later in the book, so I won't spoil it here beyond saying that a prompt has to earn its place before anything gets to act on its own.

What I will say is this. The first time we turned a repeating Friday pattern into a shift-start note instead of another chart, the manager's reaction wasn't excitement. It was mild irritation that nobody had done it sooner. I'll take that reaction over a compliment every time.

The temptation, especially once AI enters the conversation, is to jump from seeing a pattern to automating a response. I think that is usually too fast. Before a system taps somebody on the shoulder, you need to know that the signal deserves interruption. Operational teams already live in a world of alerts, messages, emails and red dots. Adding another notification is not automatically helping. Sometimes it is simply moving the friction from one screen to another.

The best prompts I have seen are quiet. They arrive rarely enough to matter and specifically enough that the person receiving them knows what to do next. That takes more understanding than building a dashboard, not less. You have to know which pattern matters, when it becomes actionable and who is in a position to change it. The clever bit is often the restraint.