GEO + SEO content

Write content that does two jobs at once: rank in traditional search (SEO) and get lifted and cited by AI answer engines (GEO/AEO - Generative / Answer Engine Optimization). The two overlap but aren't identical, and the difference is the point of this skill.

Why GEO is its own thing. Classic SEO optimizes a page to rank in a list of blue links. Answer engines (ChatGPT, Perplexity, Google AI Overviews, Gemini) don't show a list; they synthesize an answer and cite a few sources. To be one of those sources, your content has to be extractable: a model needs to find a self-contained passage that directly answers a question and is safe to quote. So GEO rewards answer-first structure, modular passages, explicit facts with attribution, and clear entities - things that aren't required to merely rank but are required to get cited.

Step 1 - Gather the brief

Before writing, get these (ask if missing, infer if obvious):

• Topic and the primary query a reader would type or ask.
• Audience and content type (blog post, landing page, FAQ, guide, comparison).
• Must-include facts / stats / sources, and the brand's voice.
• Whether this is new content or an audit / rewrite of something existing.

Step 2 - Map intent and questions

Identify the search intent (informational / commercial / transactional / navigational) and the cluster of related questions real people ask around the topic. Those questions become your headings, phrased the way people actually ask them.

Step 3 - Write answer-first, in extractable chunks

• Open with a direct answer in the first 1-2 sentences, then a short TL;DR.
• Break the body into modular passages (~40-120 words) that each stand alone and fully answer one sub-question. A model should be able to quote any single passage without needing the rest.
• Use descriptive, question-shaped H2/H3s that mirror real queries.
• Name entities explicitly and define key terms inline (answer engines disambiguate by entity).
• Back claims with concrete statistics + attribution; engines preferentially surface cited facts. Use comparison tables and lists - they're trivially extractable.
• Write quotable declarative sentences - crisp, factual, standalone.
• Show E-E-A-T: author / experience signals, real sources, current dates.
• End with an FAQ of 3-6 real questions and tight answers (doubles as extractable Q&A and supports FAQ schema).

Step 4 - Layer the SEO mechanics

• Match search intent; cover the topic to a depth that fits the query (don't pad).
• Primary keyword in the title, H1, URL slug, and first 100 words, plus semantic variants and related entities - naturally, no stuffing.
• Logical heading hierarchy; scannable formatting; descriptive image alt text.
• Suggest internal links and one or two authoritative external links.
• Provide a meta title (~55-60 chars) and meta description (~150-160 chars).

Step 5 - Deliver content + a "what I optimized" checklist

Output the finished content, then a short checklist covering: primary query + intent, GEO elements present (answer-first, modular passages, FAQ, stats, table, entities), SEO metadata (title / meta / slug / headings / links), and caveats - anything the user must verify.

Honesty guardrail

Never invent statistics, studies, or citations to look authoritative. Fabricated facts get a page filtered out of AI answers and destroy trust. Where a real number or source is needed and you don't have it, insert a clear placeholder like [STAT: add source] and flag it in the checklist for the user to fill in.

Audit / rewrite mode

When improving existing content, first note what's already working and the specific gaps (missing direct answer, buried lede, no extractable chunks, no stats / sources, weak headings, thin entities, missing metadata). Then rewrite applying Steps 3-4, and list the changes you made and why.