AI systems are becoming increasingly adept at filtering and synthesising information. For non-profit organisations and the public sector, it is therefore crucial to understand how these systems decide which sources to recommend. A recent study by Leoprd, the “Reputation to Revenue 2026” report, sheds new light on this: more citations do not automatically lead to a higher chance of AI recommendation. What counts is the quality and degree of differentiation of the evidence.
This study, reported by PRovoke Media, demonstrates that brands frequently appearing in AI responses but not chosen as the primary recommendation had an average of 9.2 citations. Brands selected as the primary recommendation, however, had an average of 8.4 citations. This suggests that the content of the information, and how it provides a clear and distinctive reason for the recommendation, is more important than simply the frequency of mentions.
What this means for non-profits and government
For foundations, charities, and government agencies, this is a fundamental insight. Where traditional SEO and discoverability focused on rankings and generating numerous backlinks and mentions, AI discoverability demands a deeper strategy. You must not only be visible but also relevant and compelling in the eyes of AI models.
AI Overviews and AI Mode in search engines like Google are changing how people find information. Users are asking longer questions and relying heavily on AI-generated answers, often without clicking through to the sources. If your organisation does not emerge as the most relevant option in this context, you miss a crucial opportunity to inform citizens, recruit donors, or spread your mission.
What does this mean for you?
It’s time to adapt your content strategy:
- Focus on Differentiating Value: Ensure your content clearly and unambiguously communicates what makes your organisation unique, which problems you solve, and why you are the best choice for the target audience. This applies to your own website as well as to press releases and other external platforms.
- Quality over Quantity: Concentrate on creating high-quality, credible sources that offer concrete answers. AI systems can use various types of evidence, such as commercial content, editorial articles, and reviews. The mix of these sources appears less important than the quality of the arguments they provide.
- Anticipate Different Questions: Bear in mind that users have different intentions. For example, an AI model might use different pieces of evidence when consumers ask “red flag” questions about potential problems compared to when they seek positive information. Ensure your content not only highlights your positive aspects but also transparently addresses any critical questions or misconceptions.
- Steer Towards a Clear AI Reputation: An “AI reputation” is not static. Building credible, differentiated, and easily interpretable digital signals helps AI to understand your organisation and, more importantly, to recommend it.
Source: PRovoke Media (Leoprd’s Reputation to Revenue 2026 study)
