Traditional per-feature AI pricing creates friction and limits adoption because customers must decide which users get access. Embedding AI everywhere solves that but risks runaway costs. The speaker proposes a middle path that balances revenue capture with cost control.
Chapters
I'll arrogantly say I think we came up with the most clever model I've seen in AI. So basically your normal choice is you either tell the customer you have to pay us for the AI thing and then you buy the separate thing. And the problem with that is then you only have a certain percentage of customers adopting it. They have to figure out like, well, Sally is going to use it more than Bob is going to use it. So why would I pay the same fee for both of them? It's like friction. It's a lot of friction in the sales motion. And then there's a lot of testing, a lot of POCs. People only have budget for a few AI things right now. So we didn't like that model. The alternative model is you just say, hey, we're just going to embed AI into everything. And then you're like, well, shoot, now my costs just explode.