The Client Calls You Shouldn't Automate

I was listening to a webinar on the state of customer success recently, and I’ve been chewing on one particular exchange ever since, because it’s the cleanest version of the mistake I see owners about to make with AI. Here’s a lightly edited version:

A CSM’s job is to take every customer from zero to ten on how well they know and use the product. What if we could do zero to four digitally? In the product, by webinar, by email. Then all of a sudden our CSMs can turn it up to 11, because now they have time for the strategic conversations.

The next panelist went further.

When people share their fear about this, the question that comes back is: do you like it? Do you like putting X and Y into a CRM? Do you like being able to tell a client where to click on a screen share? Are those the things that fulfill you? It’s all about taking some of the mundane tasks that are important for functionality, but not ultimately strategic. How do we get really smart on not spending our human time on that? And how do we then double down exactly on speaking the client’s language?

Couldn’t disagree more.

If you can automate that CRM entry, you should. But the screen share where I’m showing somebody where to click? That’s my opening to learn the client’s language. The small moments where a client tells me what they’re trying to get done are where the strategic conversation starts. The panel wants a team that speaks that language, and the plan is to automate away the only place you learn it.

When I’m having that everyday conversation with a client it’s not just click here and look there. If I’m paying attention I also find out why they’re doing it, how often, who it goes to when they’re finished, if it’s part of a bigger project, and what they did before we showed up. I don’t necessarily learn all of that in one shot. I piece it together over the course of the relationship.

Now put the owner’s hat on and multiply that across a few dozen or a few hundred clients and you have something no dashboard will ever hand you: a working picture of what your clients are actually doing with what they bought from you, and where they think the value is.

The AI-automate-it crowd will tell you the dashboard behind that zero to four experience gives you the same thing on demand. It doesn’t. It gives you usage, not intention. The “why” a client does something is more important than their doing (or not doing) something. A client running the same report every Monday morning might be your biggest fan, or might be somebody who never learned there was a faster way and is quietly getting annoyed about it. In the data those two look identical. On a call they look nothing alike. And you’ve just automated away the call in the name of “efficiency.” Strip away the “why” and value becomes usage.

Beyond losing context, the bigger loss is earned engagement. I helped them today. It was small, it took ten minutes, but they were stuck and now they’re not. Tomorrow when I ask for an hour to talk about where their business is headed and how we fit into it, they say yes. Because I earned it, not because it’s in my playbook and management expects it. That’s also where the renewal gets decided long before anybody sends an invoice.

If you own the business, you’ve never had this problem. Everyone wants to talk to the founder. The person who handles your clients after the sale doesn’t have that advantage. The only credibility they walk in with is the last conversation they had. The strategic conversation you and everyone on that panel wants them to turn up to 11 isn’t a given. They earned it at zero to four. Automate it away and you’ve taken away their ticket to the bigger conversation.

I watched a company learn this the expensive way. I was the team lead on the client service team that was, for a stretch, the sales team’s best closing argument. When a prospect asked what support was like after the sale, our reps told them to pick up the phone and call us. That was how they closed, and it worked. A new senior leadership group came in, looked at the same team, saw a cost center full of low-value zero to four interactions, and wanted it automated. Our biggest competitor heard the announcement and doubled down on their own “white glove” service model. Over the next four years the company gave up an 11 point market share lead to that competitor. Those zero to four interactions were how clients engaged with us. Take them away and we couldn’t listen the same way, and we stopped speaking their language.

In short, AI and automation are beside the point here. This is the same fight as every cost-cutting wave before it, and it ends the same way every time. Nothing on a P&L has a line for what you learned on a call, so the call shows up as a cost and whatever you learned is, at best, an intangible. Cut the call, the number gets better, and “the line goes up.” Retention is where it shows up eventually, and by then it has a competitor’s name on it.

So before you automate zero to four, ask one question, and ask it of the person who handles your clients, not your dashboard: what did we learn on those calls last quarter that the data couldn’t tell us? If the answer is something specific, keep them. If it’s nothing, automate away. If it’s “we don’t know,” you risk skipping the part where you earn the knowledge for yourself, and you know how that worked out for the park.