01 · Baseline
Run the AI baseline first
Before the session, test the same prompts in ChatGPT, Perplexity, Google AI Overviews and Gemini. Capture the exact wording each returns, and screenshot it. You want a live starting point, not a guess.
Why it matters. You cannot plan a lift without a starting point. Ten minutes of testing tells you whether the client is visible, invisible, or misrepresented in AI answers, and gives you concrete examples to open with.
Brand prompts. "What is [brand]?", "Is [brand] any good?", "[brand] reviews", "[brand] vs [competitor]". Is it described accurately? Any wrong or outdated facts?
Category prompts, no brand. "best [category] in [location]", "who are the top [category] providers", "recommend a [service] near [area]". Is the client mentioned at all? Who is?
Buyer-question prompts. The real questions a customer asks before buying. Does the client content get cited, or a competitor?
Note the citations. Which sources does the AI quote (their site, directories, reviews, press, competitors)? That is your target list.
Listen for: Wrong or missing facts about the client, competitors owning the category answers, and which third-party sources the engines trust. These three findings shape the whole plan.
06 · Content
Content and structure readiness
Answer engines lift clear, well-structured, directly-answered content. This tells you how much is new content versus restructuring what exists.
Why it matters. It tells you how much of the win is new content versus simply restructuring what already exists.
Do you have an FAQ or help section, and does it answer the real buyer questions?
Is your key information written as clear, direct answers, or buried in marketing copy?
Do you have detailed pages for each service, product and location, or thin ones?
Is important information locked in PDFs, images or gated forms a crawler cannot read?
Who creates content, at what cadence, and who signs it off?
Listen for: Clean answers win. Marketing fluff, thin location pages and gated facts all block AEO.