How to Analyze Search Results That Include AI Answers

To analyze a search results page that includes AI answers, record what appears from top to bottom: whether an AI summary is shown, which sources it cites, which brands it names, and what classic results, ads and local listings follow. Then compare those sources with your own pages. The goal is to find the content formats and sites that search engines and assistants currently trust for each query.

Step 1: Choose queries carefully

Pick 15 to 30 queries that matter for revenue. Mix types:

• Question queries: “how often should a boiler be serviced”
• Commercial queries: “best boiler service plan UK”
• Local queries: “boiler repair Manchester”
• Comparison queries: “Brand A vs Brand B boiler cover”

Step 2: Capture the page in a neutral way

Use a private browser window, set the right country and language, and note device type. Results vary by location and user, so keep conditions the same each time you check. Take a screenshot and note the date.

Step 3: Record each element

For each query, log in a spreadsheet:

1. AI summary shown? (Google AI Overview, or Copilot answer in Bing)
2. Sources cited in the AI summary, with domains
3. Brands named in the AI text
4. Ads present above or around the results
5. Local map pack, if any
6. Top organic results
7. Other features: videos, forums, “People also ask”

Step 4: Look for patterns in cited sources

After 20 or so queries, sort the cited domains by frequency. Typical patterns:

• Review and comparison sites dominate commercial queries
• Government or official bodies appear for regulated topics
• Forums and video platforms show up for “how to” problems
• Competitors’ own pages appear when they answer the question directly

Each pattern suggests a different action: get listed, publish a better answer, create a video, or earn a mention on a cited site.

Step 5: Inspect the cited pages

Open the top cited pages and note why they might be chosen: a direct answer in the first paragraph, a clear table, a date showing freshness, an author with credentials, or detailed specifics. Our guide on AI SERP optimization lists common formats.

Step 6: Go beyond the search page

Many buyers skip search and ask assistants directly. Run the same queries in ChatGPT, Perplexity and Gemini and compare sources with the search page. Overlap tells you which sources matter everywhere. See also Perplexity source tracking.

Step 7: Turn findings into a plan

• Queries with no AI summary: focus on classic SEO for now
• Queries where competitors are cited: improve your matching page
• Queries where third-party sites dominate: seek honest listings, reviews or coverage there

Repeat the analysis monthly with the same queries to see changes.

Common mistakes in this analysis

Checking from your logged-in browser, where personal history can shape results. Looking only once, when AI summaries can appear or disappear between days. Ignoring the queries where no AI summary appears, which still matter for traffic. And copying a cited competitor’s page instead of writing a better, original answer with your own facts.

Keep a simple archive

Save screenshots in a folder by month and query. When something changes, such as a new source appearing or your page dropping out, you can look back and see exactly when it happened and what else changed around that time.

How Keboost helps

Keboost automates the assistant side of this analysis. It tracks the prompts your customers ask across AI models, shows where you and competitors appear, and keeps history for trends. It then helps you create focused content for the gaps. Run a free AI visibility check to get started.