AEO strategy, playbooks, and teardowns
Short, concrete writing about what's changing on ChatGPT, Claude and Perplexity — and how B2B teams earn citations instead of fight for blue links. New posts land as we ship.
Why ChatGPT citations don't follow Google rankings
Breakdown of the signals LLMs weight vs traditional SEO — and why the overlap is smaller than most SEO tools assume.
How we probe 5 engines without burning $100 a run
Cost-attribution architecture: per-call OpenRouter logging, kind-tagged metadata, and how we keep a full diagnostic under 2 dollars.
Branded vs unbranded: the only visibility number that matters
If you're cited 80% of the time when your brand is in the prompt, you're not winning AEO. You're winning your own name. Here's how we separate the two.
The AEO checklist we run before every new workspace
32 checks across schema, content, citations and authority that decide whether the agent can even help you yet.
Prompt naturalness: why bad generated prompts tank your signal
An LLM asked 'SuperBrand review for tech' returns garbage because no real buyer types that. Here's the filter we add to kill those.
Attribution without causation: how we score action impact honestly
±14-day windows, the correlation disclaimer, and why we refuse to claim 'this article earned you 40 citations'.