His point was not that the enterprise software market would be upended immediately by AI, but rather that the pay-to-play scam of analyst-powered purchasing advice is near the end of its lifecycle.

If you've ever been part of the process, you learn quick that it's one analyst who works whatever beat your company operates in who has an extremely poor understanding of your product, the market, or where it's headed. But they'll have a new catchphrase they've dreamed up and so it's just a game of saying "yeah, sure, we do that" and then paying money to be mentioned.

I still recommend to companies that they should endeavor to be put into a Gardener Magic quadrant because it can be transformative for enterprise sales pipeline. But I always feel bad for the purchasing decision makers as non of this is good data. I agree with swyx that automated deep research will phase this whole model out, which will be a net win for both companies and customers.

I wouldn’t bet on automated deep research until they figure out a business model that gives people a throat to choke. Enterprise software is a world where it’s more important to have another human you can blame for when you fuck things up than actually making a good decision. What incentive is there for an exec to say, “well I ran a deep research and it seemed good enough to me” when their boss demands an answer as to why $VENDOR was a bad choice?