As generative search engines, ChatGPT, Perplexity, Gemini, or Claude, redefine how audiences discover culture, an institution’s visibility is no longer determined solely by traditional SEO. It now depends on its ability to be cited by AI as a reference source.

Before building a GEO strategy, it is essential to conduct an audit. This foundational step allows you to measure your institution’s actual presence in AI-generated answers, identify strengths, blind spots, and areas for improvement. The white paper GEO: How to Make Your Cultural Institution Visible in the Age of Generative AI outlines a methodology in seven key steps.

1. Test Representative Queries

Start by simulating the questions your audiences are likely to ask: tourists, teachers, journalists, researchers. For example, ask ChatGPT or Perplexity, “What are the major exhibitions in Paris this summer?”, “Who painted The Wedding at Cana?”, or “What to visit in Dordogne?”. These tests reveal the prominence of your institution in the generated responses.

2. Observe Citations and Paraphrasing

Check whether the engines mention your institution, its website, or its content directly. Are they replaced by other sources like Wikipedia or specialized press? This observation gives insights into the reliability and recognition of your data in AI corpora.

3. Assess Relevance and Accuracy

Note your institution’s position in the responses: first mention or secondary. Verify the accuracy of the information and identify any errors or approximations, the infamous “hallucinations” of language models. This critical review is essential to prevent any distortion of your image.

4. Consider Different User Profiles

AI-generated results vary depending on the context and the presumed user profile. A teacher, researcher, or foreign visitor will not receive the same answers. Testing multiple “personas” helps evaluate the coherence and accessibility of your content for each audience type.

5. Craft Specific Prompts

GEO relies on conversational phrasing rather than simple SEO keywords. Build realistic prompts inspired by the actual questions your audiences ask. This natural, contextual approach increases the likelihood that your institution will be cited in AI-generated answers.

6. Compare Generative Models

The generative search market is fragmented, each engine prioritizes different sources. Comparing your visibility across ChatGPT, Perplexity, Claude, and Gemini helps identify gaps and focus your efforts on platforms where your institution has the most potential.

7. Test with Neutral Profiles

Finally, to avoid biases linked to history or personalization, it is recommended to conduct tests using accounts without browsing history or through external third parties. This ensures a more objective view of how an AI actually “sees” your institution.

From Audit to Action Plan

A GEO audit is not an end in itself but a strategic diagnosis. It should lead to a clear roadmap:

• Revise missing or low-visibility content
• Strengthen reliable sources and structured data
• Increase mentions in media and reference platforms
• And finally, regularly track visibility evolution across generative engines

Going Further

Ask Mona’s white paper GEO: How to Make Your Cultural Institution Visible in the Age of Generative AI not only details this methodology but also presents concrete actions to implement after the audit, including data structuring, source verification, internal governance, and editorial partnerships.

‍