Recording every call, even one-on-ones

Last Wednesday I announced in the company Slack: from Q4 we collect transcripts of all calls. Even 1:1s. Even the ones with people partners.

— And what about our privacy? — the people asked.

We are not machines producing corporate data. In a one-on-one, the work stuff inevitably mixes with the personal. If everything gets recorded, we will start filtering every phrase. AI will happily pull a thought dropped by accident out of context, turn it into an "objective signal," and neatly put it at the base of an evaluation, a promotion, or a firing. We don't know who gets access to this data, how long it is stored, or whether we can delete or correct anything. We will tell less truth, or just move the important conversations to where nobody records them. 😐

— And what about efficiency? — the business answered.

For business, people are the data. We work in service, a devshop. We need to spot and promote the people who can communicate effectively, facilitate, build relationships, run retrospectives, and not be toxic. It is exactly these competencies that lead to client problems getting solved and to the client being comfortable. That is what he pays the money for. And hiring people just to fire them is dumb; business doesn't work like that.

Until AI showed up, a company under 100 people simply had no way to analyze the calls and behavior in them. And now, please: "Claude, show me the examples where this people partner uses coaching in her work." Or "find me the places where this manager was rude." And there is your material for evaluating effectiveness. 👀

By the way, right now, evaluation in the company is built on the subjective opinion of people. That opinion quietly rests on the last two months of working together, because whatever happened before that, nobody really remembers anymore. A full transcript, on the other hand, keeps the phrase together with everything that was said around it. That is exactly what nobody has at hand when the argument about context starts.

So it turns out that AI threatens to bring even more objectivity, it breaks the argument about being pulled out of context, and at the same time it climbs into everything.

And yes, the biggest fear — that I will be evaluated by AI — is the reality now. It will be.

So, dear managers, HR people, and everyone who has ever written a performance review, how do you evaluate people right now, and how much of it is just what you happen to remember? 🤔