And how AI keeps us from being intimate with our data
❤️ Let’s talk about intimacy. 🩷
Data intimacy, of course!
Human-style intimacy vs AI-style distance.
I ran a chunky dataset through Copilot in Excel and in 2 minutes it spit out a pretty clear and accurate forecast for the next 12 months.
Amazing!
Terrifying!
Work that used to take me hours now takes moments!
Dang it!
Normally, forecasting trends means aggregating search volume, social chatter, survey data, competitive content etc. and parsing it through some custom model I’ve built from my sweat and tears. It is a slow, clunky, and often demoralizing process that usually results in beautiful and gratifying work. Through that process I get to know the data. I’m working shoulder-to-shoulder with it. I notice its quirks. I can feel its shape and topography, I can sense when it’s lying. And so I can make better decisions because I’ve lived with it for a while.
Remember, the data here is the representation of our audience, their behaviors, sentiments, desires, and fears.
With Copilot? I cleaned the input, explained a custom decay model, and…that was all. Predictions acquired. Time to make a decision on it I guess.
But wait, am I only a decision-maker here, and not an analyst at all? Am I no longer a proud steward of data integrity and meaning?
Did AI just Peter Principle me into Management?
!We’ve always relied on tools to do the math. But what makes us BestFinestMarketers is our closeness to the data. It’s an intimacy with the world we’re investigating. A familiarity with the landscape, the category, the audience. And this is where the good insights come from. Not from the final output, but from the process of understanding the input. AI is fast, smart, and useful. But it puts distance between us and the world we’ve set out to understand. Distance hides nuance, blocks intuition, and ultimately sterilizes strategy. The machine doesn’t know what questions you forgot to ask. It can’t sense that something feels off. It doesn’t *care* if your dataset contradicts your brand instincts.
We should use AI. But we also need to stay in the room with the data. Touch it. Question it. Invite it up for coffee. If we hand over the hard work of intimacy with our data, are we really even doing strategy anymore, or are we just signing off on someone else’s? Why would we choose to work this way?
Look, I recommend using Copilot to run your Python in Excel. I’ve always believed smart analysts shouldn’t have to learn to code. And I can’t imagine ever going back to doing it myself. But we have to find new ways and processes to get in deep with our data.
Otherwise, what are we doing all this for?
