If it wasn't clear already, this should demonstrate how important GEO and AEO are becoming. I'd encourage founders to run their site through one of the many free checkers, such as Cloudflare's Is Your Site Agent-Ready?, to make sure AI agents can at least find and read it. If you want more detailed tracking, there are many open-source tools, including ours, Overheard AI, that track what AI answers say about you and your competitors over time. Start tracking this now. Your future self will thank you.

Is this Salesforce's path to a bigger marketing suite, like HubSpot's, or at least partly the trend in data deals for AI? In the past month it also agreed to buy Loops (AI analytics) and Listen Labs (AI customer research and simulations), and ZoomInfo added Workbounce's company knowledge to its agents. This trend of legacy SaaS integrating more data into their agent access points, and purchasing / building more AI analytics will only continue. For startups, consider what proprietary data sets, or data transformations you're building that are both defensible and valuable looking ahead.

Together these raise two trends to watch. First, the longer-term trend to consider is what will the future split between AI ads and AEO/GEO be for brand strategies? Second, right now, startups should be aware of the opaqueness of AI recommendation data. Ex. Haynes' numbers undercount AI visits, because AI often recommends a product without a link, so people Google it, or it links but strips the tag that marks the visit. At the very least, founders should do the AEO and GEO basics now and continue to ask customers how they found them.

If your rankings have shifted at all in the past few weeks, check Search Console before you change anything, since it reports Google's own data.

This story, to me, raises the larger question of the future of software. In my mind, to put it very simply, traditional software mainly provided (among other things): structured data schema and structured workflows. Does that same business model hold true in the age of AI, as the cost and barrier to building SaaS applications (especially niche SaaS applications) trend towards zero? If it does, where and for who? My feeling is that in GTM, which is rapidly trending toward more technical skill sets, a niche application like this may be challenged, since it competes with open-source projects, ours included, which already cover several of the differentiators Different claims. It's something we'll be watching as more SaaS businesses lean into distribution through APIs and MCP.