In the last month, Anthropic and OpenAI both stood up consulting arms, funded by private equity, to send their own engineers into companies and get the technology actually working.
I think this is one of the more honest moments we've had in AI.
The model is the easy part to buy, but the hard part to use.
The model is a swipe-of-the-card decision...You can have a great one by this afternoon. But you can't swipe a card for the parts that create value: clean data, real security, the judgment to know where the tech belongs, and the ability to integrate it into how your business runs.
That gap is an implementation problem, not a model problem. And implementation is custom, specific, and human before it is technical.