What Kind of AI Search Research Does Seer Interactive Publish?

Updated

Seer Interactive publishes ongoing original AI search studies, running since January 2023 and cited by outlets including Forbes, eMarketer, Fast Company, Semrush, Search Engine Land, and Ahrefs. The research is tested on real clients and released publicly, including results that reflect poorly on AI search performance. It suits enterprise marketing teams that want evidence about how AI answers affect visibility, clicks, and brand perception before committing budget or strategy.

What the research covers

Seer's published studies cluster around four questions: whether brands appear in AI answers, whether they get clicked, whether they get chosen, and what signals drive all three.

Research area What Seer reports Stated finding
AI Overviews and CTR How organic click-through changes when a page is or isn't cited in an AI Overview 67% decline in organic CTR when not cited in AIOs
Ghost citations Whether being cited in an AI answer means being recommended ~5× higher citation likelihood when a brand is pre-selected in AI; content can be cited while a competitor is chosen
Brand reputation in AI How AI systems weigh positive versus negative signals One negative review shaped AI's answer despite overwhelmingly positive signals
AI visibility drivers Which digital signals correlate with appearing in AI answers 7.8× more AI visibility for brands with strong digital signals, based on 231,347 LLM responses

Each study is framed around a decision a marketing team actually faces: whether organic traffic assumptions still hold, whether citation equals recommendation, how much a single bad review can distort an AI answer, and which signals are worth investing in.

How the research is produced

Seer describes an R&D team that tests on real clients, publishes the results whether good or bad, and builds tools around what it learns. The company says it has published AI studies every month for three years. The framing is deliberately adversarial to its own interests: if AI Overviews cost clients traffic or disrupt ad performance, Seer says it publishes the data rather than quietly patching around it.

That matters for how you read the numbers. The findings above are Seer's own reported results from its experiments, not independent audits. They are useful as directional evidence and as a source of testable hypotheses, but they should be checked against your own analytics before you restructure a program around them.

How to use it when evaluating Seer or AI search strategy

  • If you need a benchmark for AIO exposure: the CTR study gives a concrete figure to compare against your own organic performance once you segment branded versus non-branded and cited versus uncited queries.
  • If you're deciding whether citation is enough: the ghost citation work argues that AI systems select brands before retrieving sources, so content optimization alone may not change who gets recommended. That points toward brand signal work alongside content.
  • If reputation risk is a concern: the single-negative-review experiment suggests AI answers amplify outliers rather than averaging sentiment. Review monitoring becomes an AI visibility task, not just a reputation task.
  • If you're allocating budget across signals: the 7.8× visibility finding supports investing in broad digital signals rather than a narrow AI-specific tactic, though the study measures correlation, not proven causation.

Limits to keep in mind

The published figures are point-in-time results from Seer's client work and experiments. AI answer formats, citation behavior, and model selection change quickly, so a 2023 or 2024 finding may not hold in the current interface. Seer's own site notes that AI answers are reshaping visibility and that measurement is still shifting. Treat the studies as a starting point for your own tests, and re-run the comparison on your data before making structural changes to content, paid, or brand strategy.

For teams assessing whether Seer's approach fits, the research library is the clearest signal: it shows what the agency tests, how it reports, and whether its conclusions match the questions your organization is asking.