Why ChatGPT Isn't Saving Your Team Time
Updated: Jul 26
I find it interesting how quickly "AI transformation" has become synonymous with "everyone has ChatGPT now."
Those aren't really the same thing. If anything, ChatGPT reminds me of the early days of Excel. It made millions of people individually more productive, but it didn't fundamentally change how businesses operated. You still had the same approval processes, the same meetings, the same handovers and the same bottlenecks. You just had nicer spreadsheets. I wonder if we're seeing the same thing happen again.
I know plenty of people who use ChatGPT every day. Not going to lie, I do too. In fact, I'd hate to work without it now. But if you asked me whether simply giving everyone ChatGPT transforms a business, I'd say probably not. It transforms individuals first. Businesses are a different beast entirely.
The interesting part isn't whether one person can write an email faster. It's what happens after they hit Send.
Does someone else need to review it?
Does another team need to approve it?
Does somebody copy the same information into Salesforce because the systems don't talk to each other?
Does the customer now wait three days because the next person in the chain is on leave?
ChatGPT hasn't made any of that disappear. That's why I spend surprisingly little time talking about AI when I first meet a client. I'm much more interested in hearing someone tell me what happened yesterday. Not the official process. Yesterday.
"Walk me through the last client that came through."
That's usually where the interesting stories are. People almost apologise as they describe them.
"Oh... then we export it into Excel..."
"This bit's a little clunky..."
"We have to send it to finance because..."
"We actually check it twice..."
They assume these things are normal because they've been living with them for years. That's where I start getting excited. Because by the end of that conversation we've usually forgotten we were supposed to be talking about AI. We're talking about the business instead.
And once you've untangled that, the AI part is often the easy bit.


Comments