AI search field guide

Worked example: interpreting an AI visibility change

Illustrative example only: a team tracks 20 questions across two collection surfaces. The numbers below teach the method and do not describe an InfiniteRanks customer.

InfiniteRanks editorial team

Record the first window

Suppose the team requests 80 checks over two days. Sixty-four return eligible answers and sixteen fail. The brand is mentioned in sixteen eligible answers, so mention rate is 25%. Citation metadata is complete for 48 eligible answers, with the owned domain cited in six, giving a 12.5% citation rate. Report the 64/80 collection coverage separately.

Compare the next window

Suppose the next window has 72 eligible answers and 24 mentions, a 33.3% mention rate. That is an increase of about 8.3 percentage points. Before claiming progress, confirm the same question cohort, markets, methods and entity definitions. If new easy branded prompts were added, compare the unchanged subset first.

Inspect the answers behind the change

Read the questions that gained or lost a mention. Check whether the wording is favorable, neutral or qualified and whether citations support the relevant claim. A rising mention rate could include cautionary references. A source link and a brand mention are separate events and should remain separate in the narrative.

Write a defensible conclusion

A supported statement is “the owned brand appeared in more eligible answers in this unchanged sampled cohort.” An unsupported statement is “the campaign increased revenue by 33%.” List page changes and other events, then explain that a before-and-after observation alone does not isolate their causal effects.

Sources and further reading

Sources explain the underlying standards. They do not endorse InfiniteRanks or guarantee a ranking outcome.

Be part of the conversation.

Explore a model response with your own provider key.