If your brand matters in AI-generated answers, where people talk about you matters too — and not always where you expect. An Ahrefs analysis of 10,000 brands and roughly 260 million non-branded AI prompts found that Reddit threads correlate more closely with how often brands appear in AI responses than any other off-site mention type. That finding should change how PR, comms and SEO teams prioritise monitoring and outreach.

What the data shows

Ahrefs compared how frequently brands appear in AI outputs across three surfaces (ChatGPT, Google’s AI Mode and AI Overviews) with seven types of off-site mentions: Reddit threads, domain-level mentions, webpage mentions, traffic-weighted webpage mentions, news articles, YouTube mentions and TikTok mentions. The study measured Spearman correlations between each mention type and AI mentions.

  • Reddit mentions produced the highest correlation with AI mentions (≈0.75).
  • Domain mentions, webpage mentions and traffic-weighted webpage mentions each correlated about 0.70.
  • News article mentions correlated ~0.67; YouTube ~0.64; TikTok ~0.60.
  • Brands with 10× more webpage mentions typically received ~3.7× more AI mentions; a 10× increase in Reddit mentions mapped to ~5.5× more AI mentions.
  • ChatGPT showed slightly weaker correlations with off-site mentions than the other two AI surfaces in the study.

The relationship is clear but not linear: increases in off-site mentions tend to boost AI visibility, but with diminishing returns. Tenfold web coverage does not equal tenfold AI visibility — a sustained and broad spread of mentions matters more than a handful of high-profile placements.

Why Reddit tracks AI visibility more closely

Two factors help explain Reddit’s stronger correlation. First, Reddit content is both widely crawled and licensed: major AI systems and search providers have ready access to Reddit posts and comments, increasing the chance that Reddit discussions inform AI outputs. Second, Reddit often hosts community-driven, comparative and recommendation-style conversations that describe brands in contextual ways — the type of textual usage language models learn from when producing answers.

That said, correlation is not causation. Reddit’s signal may partly reflect that brands with active, discussion-driven followings are inherently more likely to be represented in conversational AI outputs.

How marketers should respond

Treat press hits and big placements as valuable, but not sufficient. The data suggests a practical shift in priorities:

  • Build breadth, not only marquee placements. A wider spread of mentions across many sites and domains correlates with greater AI visibility than a small number of high-traffic mentions.
  • Monitor and engage relevant community platforms. Where it fits your audience, track Reddit threads and participate in topic-specific discussions that naturally surface your brand in contextual recommendations or comparisons.
  • Plan for scale and time. Because returns diminish, meaningful changes in AI visibility typically require sustained increases in the number of mentioning sites and months of measurement.
  • Measure across AI surfaces. Correlation strengths differ by platform; monitoring multiple AI outputs provides a fuller picture of visibility trends.

Practical measurement tips

Ahrefs used its Brand Radar and Content Explorer to collect AI mentions and off-site references. For your measurement program, apply these principles:

  • Separate branded and unbranded AI mentions. The study excluded prompts that directly named brands to focus on organic references — a useful distinction when tracking visibility growth.
  • Combine social listening and web-mention tracking. Use tools that capture Reddit, YouTube and TikTok activity alongside web mentions and domain counts so you can correlate changes over time.
  • Segment by brand size and search demand. Brand size explains part of the relationship; compare peers within similar search-volume bands to detect independent effects from off-site mentions.
  • Run controlled experiments and long-term time series. Because the analysis is correlational, set up outreach bursts to many niche sites or coordinated community engagement and measure the impact on AI mentions over months.

Note the study focused on U.S. brands and excluded short or common names, so detection rules may require adaptation for smaller or generically named companies.

What to watch next

Source access matters. If models or platforms change which sources they license or prioritise — for example, tighter restrictions on licensed data or shifts in crawl priorities — the mention types that best predict AI visibility could change. For now, the actionable takeaway is straightforward: invest in a broad, sustained presence across many domains and community platforms, and put Reddit on your monitoring list. Track progress across multiple AI surfaces and treat improvements as a medium-term project rather than a quick win.