I'm curious what you mean by "AI product"? Do you mean a product that using AI, or a product used by AI? For the former, it's the same problem as it's been forever: distribution. That can be through SEO, AEO, outbound sales, advertising, etc, etc. I do think that AEO is something you should be optimising for. At the company I work for, we've seen inbound go from virtually nothing to top three in the last 6 months.
For an expert, this will likely be a wash (at best), but I think getting AI assistance on the analysis of the data is really useful for people without a lot of experience with web analytics. The really interesting stuff comes when you combine it with other context though, adding Google Search Console, Git commits, Apollo for leads, etc.
When you're building B2B2C products, your relationship is with the customer, but all the growth and potential comes from the user, someone you might never get to know. Some thoughts based on my experiences.
Wrote something on how you can differentiate your products by focusing on fixing where they break at the edges. This is where startups can really win if they do a good job.