We also run our own research at Findabl.

There are a lot of questions about AI search right now and not a lot of good answers. So when we see a question we think matters, we study it.

We've looked at things like how often AI recommendations change, how often brands actually show up in unbranded answers, where the citations are coming from, and how different the answers are across ChatGPT, Gemini, Claude and Perplexity.

We publish the results because I think we all need more actual data about what is happening in AI search.

This is a big part of what we're trying to do with Findabl. Not just build another tool, but use the data to understand what's actually happening.