GEO
The problem: Google is not the only place people “search” anymore
A prospect asks ChatGPT or another assistant for a short list of options. The answer names three brands, maybe with pros and cons, and yours is not on it. That is not a “ranking” problem in the old sense; it is a presence and comparison problem in generated answers.
Generative engine optimization is a different scoreboard: mentions, recommendation framing, and who gets used as a reference when people ask in natural language. Invention Novelty runs real prompts, stores responses, and lets you see how you stack up against the brands you care about, so you are not flying blind in the new default.
What is broken today
Classic rank tools lie by omission
You can rank “fine” in ten blue links while your category is decided in a paragraph of AI output. If you are not monitoring mentions and framing, you are planning from half the data.
Vibes, not a loop
“Be more present in AI” is not a ticket. You need runs you can compare: prompts, competitors, and a paper trail. That is the GEO workflow we built around, not a one-off screenshot.
How we approach it
Define brands and competitors, run a scan, and read the answers like a product team: who gets cited, how you are compared, and where the next content or PR move should be. It is the same workspace mindset as the rest of the product: AEO and organic structure still matter, but they are not a substitute for knowing what the models say out loud.