
This is the operational summary of our AEO series — twenty actions we run for our own brand and our clients’, ordered so the highest-leverage work comes first. Score yourself one point per item genuinely done; in our monitoring, brands above fifteen points show up in AI answers dramatically more often than brands below eight.
Entity foundation (1–6): ① One consistent name, address, and description everywhere — site, LinkedIn, GitHub, directories; fix every variant. ② Organization + WebSite JSON-LD with sameAs to every real profile. ③ A company LinkedIn page that restates your key facts. ④ Person schema and a bio page for at least one real, searchable leader. ⑤ Profiles on the review directories your buyers trust (for agencies: Clutch, GoodFirms) with verified reviews. ⑥ A public GitHub or equivalent artifact trail if you sell engineering.
Content that answers (7–12)
⑦ One canonical page per money question — not five fragments. ⑧ Published pricing or honest ranges; "contact us" is invisible to a model answering "how much does X cost". ⑨ A definitional pillar for your category, answer-first. ⑩ Comparison content with real criteria, including alternatives to you. ⑪ Case studies with concrete numbers and named outcomes. ⑫ Dated claims refreshed on a schedule — stale numbers get demoted by freshness-weighted retrieval.
“Above fifteen points, brands show up in AI answers. Below eight, they are relying on luck.”
Machine legibility (13–16)
⑬ FAQPage, Service, and Article schema on the pages that carry those shapes, validated and in sync with visible content. ⑭ Full text server-rendered — no content locked behind JavaScript execution or login walls. ⑮ robots.txt explicitly permitting the AI crawlers you want (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) — blocking them is opting out of the answer. ⑯ An llms.txt stating canonical company facts for systems that read it.
Measurement and compounding (17–20)
⑰ A frozen prompt panel — 30+ buyer-realistic questions — run monthly across the engines that matter. ⑱ Track mention rate, accuracy, sentiment, and citation share as trends, not screenshots. ⑲ Publish at least one piece of original data per quarter — the only content class engines reliably cite rather than paraphrase. ⑳ Assign an owner. AEO is a compounding practice with a feedback loop, not a one-time project; unowned, it decays like everything else.
Run the checklist quarterly and let the score trend tell the story. If you want the measurement layer without building it — that is exactly what Sourceable does, continuously, across four engines. Either way: start with items one through six. Entity before content, content before tactics.
Emperor Brains Engineering
Emperor Brains LLP


