I subscribe to roughly thirty newsletters about artificial intelligence. Most of them are terrible. They rehash the same announcements, use the same breathless language, and provide the same shallow analysis that could have been generated by the very tools they are writing about. I keep subscribing because occasionally — rarely — one of them publishes something genuinely useful.

Six months ago, one of them did. It was not a trend piece. It was not a prediction about what AI would do by 2030. It was a 1,200-word case study about how a four-person accounting firm in Ohio used AI to cut their tax preparation review time by 60%. The piece included the specific tools they used, the exact workflow they built, the mistakes they made along the way, and the actual dollar figure they saved in the first quarter.

I read it twice. Then I forwarded it to eleven people. It was the most useful thing I had read about AI in six months, and it had nothing to do with the technology itself. It had everything to do with the specificity.