How to Check if ChatGPT
Recommends Your Brand

Why this matters before we start

When a buyer asks ChatGPT "what's the best [your category] software for [use case]?" — your brand either appears in the answer or it doesn't. Unlike Google, where you might rank #8 and still get some traffic, an AI answer that doesn't include your brand produces zero exposure for that buyer. No second page, no "also check out" section.

Checking your current citation status is the first step in any AEO strategy. It establishes your baseline — and tells you whether you have a problem at all.

Step 1: Build your 10 test queries

Don't just ask "what are the best [your product] tools?" once and call it done. AI responses vary significantly based on how a query is framed. Build a list of 10 queries covering the different ways your buyers ask about your category:

The pattern you're looking for: which query types show your brand, and which don't? The ones where you're absent are your highest-priority AEO gaps.

Step 2: Run the queries on ChatGPT

Open chatgpt.com and run each query fresh. A few things to note:

• Use a new chat for each query. ChatGPT's conversational context can influence subsequent answers in the same thread. Each query should be independent.
• Don't be logged in as an employee. ChatGPT may personalise responses based on account history. Use an incognito window or a clean account.
• Record exactly where you appear. Note your position (#1, #2, #3, absent) and which competitors appear before you.
• Run each query 2–3 times on different days. ChatGPT responses are non-deterministic — a single test can be misleading. Consistent absence across multiple tests confirms a real gap.

Step 3: Repeat on Perplexity, Gemini, and Claude

ChatGPT is not the only AI engine that matters. Buyers use four major engines, and citation presence varies significantly across them. A brand that appears consistently on Perplexity might be absent on ChatGPT, and vice versa. You need to check all four:

Run your 10 test queries on each engine and record the results in a spreadsheet. You now have a 10×4 grid of citation positions (or absences) — your AEO baseline.

Step 4: Interpret what you find

There are four common patterns and what they mean:

The limits of manual checking

Manual checking is a good starting point, but it has real limitations for ongoing AEO management:

• AI responses are non-deterministic. The same query can produce different answers in different sessions. A single manual check is a snapshot, not a measurement.
• It doesn't scale. Checking 10 queries across 4 engines, run multiple times, is 40–80 manual tests. Every week. This isn't sustainable.
• You can't track trends. Without consistent position data over time, you can't tell whether your changes are working or whether external factors (competitor improvements, model updates) are affecting your position.
• You get no competitive context. Manual checking shows you where you appear but not why competitors appear before you or what specific actions would close the gap.

The automated alternative: Surfedo's free scan

Surfedo automates the entire process: you enter your domain and key queries, it scans all four AI engines simultaneously, and returns exact #1/#2/#3 position rankings for every query on every engine. The free scan runs the first check in 60 seconds — no card required.

Beyond the baseline scan, the paid plan runs weekly automatic rescans and tracks position history so you can see whether your AEO work is moving the needle — or whether a competitor is closing in. It also generates the content you need to improve positions: FAQ pages, schema markup, llms.txt entries, comparison pages.