The answer changed.
What changed with it?
Compare the wording, claims and sources before calling it progress.
What the comparison supports
The inputs changed. Compare again before using or downloading this result.
| Observation | Earlier | Later | What to check |
|---|
Sources added
Sources removed
A change is a lead
A stronger recommendation or a new source gives you something to investigate. It doesn't tell you what caused the change. A pair of answers is a fairly small sample to hang a victory speech on.
Keep the record with the answer
Keep the excerpts, source URLs and conditions together. Future you will want to know whether the question changed before the website did.
Your entries stay in this page. Export the record before closing it; nothing is saved automatically.
How to read it
Start with the conditions
Use the exact same question. Keep the engine, mode, language and market as consistent as you can. Record search use when it's visible. An enabled search switch doesn't prove the answer used it.
Read the sentence around the name
This tool checks for the business name as written, ignoring case. It won't recognise aliases or decide whether the answer recommends you. That part needs your judgement, which is why the fields are yours to fill in.
Keep citations separate from claims
The source comparison uses the URLs you enter, with duplicates and fragments removed. A URL beside an answer isn't proof that it supports every claim. Open the source and check the passage that matters.
Repeat the check
If the difference matters, run the same question again and keep the less flattering answers too. A useful record includes the answer you expected, the one you didn't, and the runs where no answer appeared.
Own-site citations use an exact hostname match, ignoring âwwwâ. A different subdomain is counted separately. Source pages aren't fetched or checked by this tool. It compares the excerpts and URLs you supply.
Take the next check further
These are tools I work on and research I publish. Pick the part that helps with the question in front of you.
Beamtrace
I contribute to Beamtrace at Elfsight.
- Track search visibility
Follow whether the brand appears across a repeatable set of questions.
- Track positions in answers
Look at where a brand appears when the answer presents an ordered list.
- Compare competitors
See which alternatives appear alongside the business you're checking.
- Follow brand mentions
Inspect how a business is named and discussed across answers.
- Investigate cited sources
Look into the sources attached to answers and where they lead.
- Check how the brand is described
Follow the claims and characterisations that shape an answer's impression.
BLURSOR
My independent tools and research publication.
The AI Readiness Audit examines access, server responses, raw content and the rendered page. Useful when the problem might start before an answer is written.
Research worth opening
- Query language can select the market
A reason to record language and location when comparing local answers.
- A rewrite can reduce answer visibility
Why a cleaner-looking page still needs a check under the study's conditions.
- Search use and publisher referrals
Retrieval, citation and a reader's visit are different things to measure.
Sources & notes
I work on SEO and content pipeline automation at Elfsight and contribute to Beamtrace, the search visibility tracker we're building there. BLURSOR is my independent project. Those connections explain the links above; this is an independent tool under my name.
The example is fictional and illustrates how to keep a comparison, rather than reporting a test I ran. Recommendation and tone are entered by the reader. A name match, a source count or a pair of answers doesn't establish a ranking change or explain its cause.
The BLURSOR articles discuss published studies. You can inspect the original papers on language and local market selection, body-copy rewrites and search mode and publisher referrals. Read the setup before applying a finding to your own checks.