Can you teach a machine your gut?

Happy Sunday!

I don't sit in on most client calls anymore. I listen back to the recordings after the fact, usually on a walk, sometimes the next morning with coffee. But every now and then I'll sit on a live one, usually when there's a new account manager and I want to see how they handle the room.

Here's the thing nobody warns you about. When you've done something for five years and you finally watch someone else do it, you spend the whole call screaming internally. It's not that they're bad at it. They're sharp, and a lot of them came in with serious consulting backgrounds.

The part where I have to sit on my hands

There's a moment that goes by where a founder says "we tried email and it didn't really work," and every instinct I have wants to grab the wheel. Because that sentence almost never means what it says. It usually means "I got burned once and I'm a little scared to spend money here again."

The account manager answers the surface question. The right answer speaks to the fear, not the words. And the hard part isn't hearing it — five years taught me to hear it. The hard part is choosing not to jump in.

That's a choice I make on purpose, and it costs me something every time. The read is so nuanced: the pacing, the pause before someone answers, the words they pick, the half-second of hesitation that tells you the real thing is underneath. I can hear all of it and still make myself stay quiet.

And I still couldn't write down the rulebook for any of it if you asked me to.

So I'm building the rulebook anyway

I've been building something that gives each call the kind of read I'd give it if I were sitting in. Sentiment, word choice, the gaps between words, the pacing.

After every call, the account manager gets feedback, whether or not I was in the room. The whole point is that my attention shouldn't depend on whether I happened to be on that particular call. Everyone should get the read, not just the ones who got me live.

The reason I hold back

Here's the lesson I had to learn the hard way, and it's the whole reason I built it the way I did. Saving someone time doesn't actually work.

When I jump in and catch the mistake for them, I rob them of the rep. They never feel the bad pacing land in real time, so they never learn to feel it coming. They have to fail, and I have to let them, for any of it to stick.

That's why the feedback comes after the call, not during. That quiet I make myself keep in the room — this is what it was always for. Let them sit in it live, then give the correction once it's over. The whole thing is basically the lesson I had to learn as a coach, turned into software.

The part I haven't figured out

Now here's the open question I genuinely don't have an answer to. Can I actually capture the thing that makes me good in that room, or is the screaming-internally part exactly the thing a piece of software will never catch? I don't know, and I'm shipping it anyway, partly to find out.

What's funny is I've noticed the team starting to ask "what would Raymond flag here" when I'm not around. I've become a little reference point in the third person, which is a strange thing to watch happen.

Five years of calls taught me something I can barely explain, and now I'm finding out whether you can hand that to a machine, or whether some of it only ever lived in the room.

I'll let you know what I find.

- Raymond