Your call just got scored by AI. You won't see how.

A customer service agent hangs up the phone. Somewhere in the background, software has already scored that call. The agent cannot see the criteria. They cannot see the number. They just know, eventually, that it fed into a review.

That's the picture described in a recent piece from Social Europe, looking at how AI systems now score customer service calls in full, without the people being scored getting access to the rules behind the score. Not a sample of calls. Every call. And the logic behind the grading stays with the software, not with the worker being graded.

If you've worked in a call centre, this won't surprise you much. Quality monitoring has existed for decades, usually a supervisor listening to a handful of recorded calls a month. What's changed is the scale and the silence. Every interaction gets scored now, by a system whose rules nobody on the floor has ever read.

For a customer service worker, that's a strange position to be in. You're being measured constantly, on something you can't study, improve against, or challenge. Was it tone? Talk time? A word you used or didn't use? You don't know, so you can't learn from it. That's not feedback anymore. It's just a verdict.

It would be easier to accept if the people building these systems had it figured out. They don't. A separate HR Dive survey found that very few chief HR officers feel confident they can actually manage AI-driven change in their organisation. The main barrier respondents gave wasn't cost or technology. It was that they can't predict how AI will actually affect the workforce it touches. The people buying and deploying scoring tools are, by their own admission, working somewhat blind too.

So you end up with two groups in the dark at the same time. Workers don't know how they're being judged. Employers don't fully know what their own tools are doing to the people who use them. Nobody designed that on purpose, it's just what happens when a system gets rolled out faster than anyone can explain it.

There's a reason this keeps coming up in public debate rather than staying a back-office IT issue. A recent poll cited by Social Europe found that voters want AI regulated specifically for its effect on jobs, not only for broad safety concerns. People aren't worried about AI in the abstract. They're worried about AI deciding, quietly, whether they're good at their job.

If you manage a customer service team, the practical question isn't whether to use monitoring software. Most teams already do, in some form. The question is whether you could explain, in plain language, to the person being scored, what the score is actually based on. If the honest answer is "not really," that's worth fixing before someone else asks the same question on your behalf.

We list customer service roles at Will Werk, and this is the kind of detail job seekers increasingly ask us about before they apply: not just the pay, but how the work actually gets judged day to day.