The hidden influence layer in AI search is Reddit & YouTube
For years, search strategy was built around a fairly simple assumption. If your site ranked, your brand won attention.
But today, AI-generated answers are not built solely from owned media. They pull from a wider ecosystem of communities, creators, tutorials, reviews, comparisons, and conversations happening across the internet.
To understand that shift, DEPT® tested realistic consumer prompts across major AI platforms and mapped which sources were cited. The findings are directional, but a pattern emerged: Reddit and YouTube were the top-cited sources again and again.
- Reddit emerged as a source of real human context, peer validation, and discussion-led authority.
- YouTube helped users move from curiosity to confidence through tutorials, explainers, troubleshooting guides, and expert-led content.
Together, they point to a bigger shift in search where owned media still matters, especially for foundational education. But as consumers transition into conversational searches that mimic real-world interactions, AI systems are pulling from Reddit and YouTube for human validation.
The old search funnel has fragmented
One of the clearest signals in our analysis is that AI models do not rely on a single source type throughout the journey. They pull from different platforms depending on the prompt.
For upper-funnel informational prompts, owned and editorial content still matter. But as prompts become more investigative or transactional, user-generated content becomes far more influential.
For example, on a research project for a cosmetics and wellness client, our team examined ingredients-led beauty prompts, in which Reddit was cited in 101% of responses across the sample, indicating that multiple Reddit sources often appeared in a single answer.
YouTube appeared in 82% for the same sample of prompts. For bottom-funnel prompts specifically, YouTube was cited nearly 95% of the time.
This means that if your GEO or AI search strategy is centered solely on optimizing web copy, Schema, and technical elements, you are optimizing only one layer of what the model sees. The broader layer, the one that shapes confidence, recommendation, validation, and purchase, often lives off-site. This layer is also referred to as the external narrative and is comprised of a wide range of external platforms and sources as broad as the internet.
-
101%
Beauty prompts in which Reddit was cited
-
82%
How often YouTube also appeared
-
95%
Bottom-of-funnel citations on YouTube
Reddit is where AI finds conviction
Reddit’s role is bigger than many brands assume. It is not just a source of chatter but also a source of cited perspectives and human validation.
Across our audits, Reddit repeatedly emerged as one of the most frequently cited UGC sources in AI answers. It was especially influential for top- and mid-funnel questions, where people want opinions, comparisons, routines, and plain-English explanations from other people rather than polished and biased brand messaging. Text-based discussion posts were particularly important, and more than half of the cited Reddit content included brand mentions.
Unlike fast-moving social formats, Reddit also has shelf life. In the research, LLMs frequently cited archived threads ranging from six months to more than seven years old. Strong threads do not just spike and disappear; they compound.
““LLMs frequently cited archived threads ranging
from six months to more than seven years old.”
”
For brands, that means community contribution and participation can translate into long-term discoverability and recommendations in AI platforms. If your brand is absent from the right advice threads, myth-busting discussions, and niche expert communities, LLMs will still answer the question—but without your brand being mentioned.
YouTube is where AI finds validation
If Reddit helps shape perspective, YouTube helps close confidence gaps.
Our AI search audits found that YouTube becomes more influential as users move lower down the funnel. It is where people go when they want more than a claim. They want to see the routine, hear the explanation, compare alternatives, and judge credibility for themselves. That is exactly why YouTube becomes so important in AI-assisted decision-making. And it is not just any YouTube content that gets reused.
Creator-led channels accounted for 93% of cited YouTube content. Long-form videos accounted for 94% of cited videos. One of the most useful findings was that more views did not necessarily mean more citations. Some of the highest average citation rates came from videos under 1,000 views and from videos in the 10K–100K range.
That challenges one of the industry’s assumptions that AI visibility follows the same logic as social virality. More views = more AI citations.
AI models seem to reward usefulness, structure, specificity, and authority more than scale alone. A clear how-to, a doctor explainer, or a creator review with strong metadata and a focused answer can outperform a much more viral but less relevant video.
What brands still get wrong in AI search
A lot of brands are still over-investing in what is easiest to control and under-investing in what is actually influential.
Too many brands still:
- Optimize pages but ignore creator ecosystems
- Publish product claims, but avoid community participation
- Chase reach instead of retrieval
- Mistake views for authority
- Build content calendars without asking whether those assets are likely to be cited, paraphrased, or used as validation in AI results
- Fail to use UGC platforms for business and audience insights
The result is that AI visibility shifts toward publishers, creators, forum threads, and competitors that answer the real question more credibly or naturally.
A modern GEO strategy has to build three layers at once
You won’t improve AI visibility using owned media alone. The strongest strategies incorporate the external narrative – combining owned content, expert validation, creator partnerships, editorial presence, and social/community participation.
- Owned narrative: Structured, search-led, and brand-aligned content that answers foundational consumer questions clearly.
- Community intelligence and participation: Active listening and credible contribution in places where real recommendations happen, especially Reddit and online forums.
- Creator and expert validation: Long-form, expert, and creator-led content on platforms like YouTube that helps consumers compare, understand, and trust. This can be amplified by earned media (PR, publishers, and third-party review platforms), which provide broader authority signals that LLMs can use to validate your brand.
On YouTube, think less like an advertiser and more like an educator. How-to guides and listicles represent one of the biggest citation opportunities, together making up nearly half of the videos cited by AI platforms. That makes creator PR especially important. Prioritize credible experts, comparison videos, tutorials, and practical explainers.
The strongest YouTube strategies combine owned education with external validation. Category experts and trusted creators help prove that products deliver beyond marketing claims. And because many viral fan-favorite brands still lack strong, search-optimized channels of their own, there is a clear opening for brands that can pair community validation with a dedicated, science-backed educational hub on their owned channels (e.g., their websites).
On Reddit, start by listening. Reddit is a powerful business intelligence tool for understanding real consumer language, concerns, myths, comparisons, and product feedback.
Some brands may go further by building their own subreddit community, but that is the exception, because activating on Reddit is more resource-intensive than on other social platforms. For most, the better path is credible participation. For example, using employees or expert profiles to answer questions naturally, clarify misinformation, and add value without sounding corporate. AMAs with business leaders, scientists, product developers, or other experts can also create durable content that AI may continue to retrieve over time.
AI search is more than a content problem. It is a retrieval problem. You must be able to connect owned content, community proof, and expert validation around the questions consumers are actually asking.
