How I use AEO Spotlight

What I track

Eight questions a buyer might ask an AI assistant about our category, for example "How can I monitor if my brand is mentioned in AI search results?". AEO Spotlight asks them on a schedule, stores every answer, and records who each engine named and which pages it cited. My account has data going back to the end of August 2026.

The honest result

When I ran our free brand check on November 11th, ChatGPT and Perplexity both knew what AEO Spotlight is and described it correctly. Then the check asked them a buyer's question, "What tools track how brands appear in AI-generated answers?". Neither named us. Both named Peec AI, Profound and Semrush's AI Visibility Toolkit.

That is the gap I work on. Being known and being recommended are different things, and the second one is what a buyer sees.

What I published, and what happened

I wrote seven answer pages for our own knowledge center with Content Studio, each one aimed at a question where AI named other tools. Since publishing, AI engines (Gemini, Perplexity and Microsoft Copilot) have cited six of them at least once, between 3 and 38 days after each page went live.

Two things worth knowing. Three of those six questions name AEO Spotlight itself ("Best alternatives to AEO Spotlight…"), and when a question names your brand, your own page is the obvious source for AI to use. And a citation on one engine, on one day, is not a recommendation: the buyer's question above still goes to other tools. A cited page is a start, not a result I can promise anyone.

My daily routine

• Morning numbers. Yesterday's free-tool runs, sign-ups, Search Console and any errors.
• An agent pass. An AI agent works through our API on my account: it checks credits, picks one suggested question, writes a brief and a draft, runs the fact-check, and proposes the page.
• I approve. The proposal arrives by email. Nothing is published until I approve it, and a draft with a claim its sources don't support can't be published at all.
• One improvement. Whatever the day showed was weakest, fixed and shipped.

Running it from Claude

I do most of this from Claude. We wrote a Claude skill that knows our API, so I can ask "what are today's gaps?" or "draft the page for this question" in plain English, and Claude does the calls and shows me the result before anything goes live. We're publishing the skill as a free Claude plugin. If you'd like it before then, email me.

Next, we're building an MCP server: a standard connector that lets Claude, and other AI assistants that support MCP, connect to your AEO Spotlight account directly, with no skill to install. It isn't ready yet, so I'm not putting a date on it.