About us
Which Sources Actually Influence AI Brand Mentions? We Analyzed 371K+ AI Responses to Find Out
We tracked 21 cybersecurity companies across 371,000+ AI-generated responses and 5,056 third-party citation sources to answer a simple question: when AI platforms like ChatGPT, Gemini, and Perplexity mention a brand, what sources are driving those mentions?
The results challenge a lot of assumptions about how AI visibility works, what marketers should be measuring, and where to invest their time.
How We Measured This
Using our AI Visibility Platform, we monitored AI responses across the full buyer journey, from top-of-funnel awareness queries to bottom-of-funnel purchase-decision prompts. For each AI response, we tracked two things:
- Which third-party sources the AI cited in its response
- Whether the AI mentioned the brand we were tracking
By cross-referencing these, we calculated a "mention rate" for each source: when the AI cites this source, how often does it also mention the brand? We then compared each source's mention rate against the stage baseline to calculate lift - the percentage-point difference above or below average.
The baselines across all sources:
| Funnel Stage | Baseline Mention Rate | AI Responses Analyzed |
|---|---|---|
| Top of Funnel (Awareness) | 7.9% | ~137,000 |
| Middle of Funnel (Consideration) | 43.0% | ~165,000 |
| Bottom of Funnel (Decision) | 73.5% | ~69,000 |
All 5,056 sources in the dataset are third-party domains. Brand-owned domains were excluded from the analysis.
Finding 1: Top-of-Funnel Prompts Are a Measurement Trap
The average brand mention rate for top-of-funnel queries is just 7.9%.
This is significant because a lot of marketers have started measuring their AI visibility by testing generic awareness prompts like "What is endpoint security?" or "Explain zero trust architecture." These prompts feel like a natural place to check whether your brand is showing up in AI. But the data tells a clear story: AI almost never mentions specific brands at the awareness stage.
The most-cited sources at top-of-funnel are generic, high-authority platforms:
| Source | Citations | Mention Rate | Lift |
|---|---|---|---|
| reddit.com | 12,886 | 9.5% | +1.6 pp |
| youtube.com | 10,949 | 10.2% | +2.3 pp |
| linkedin.com | 9,830 | 4.9% | -3.0 pp |
| en.wikipedia.org | 5,408 | 11.7% | +3.8 pp |
| medium.com | 4,123 | 4.2% | -3.7 pp |
| microsoft.com | 3,432 | 7.7% | -0.2 pp |
| ibm.com | 3,420 | 2.5% | -5.4 pp |
Notice a pattern? The sources AI cites most at this stage are content platforms (Reddit, YouTube, Wikipedia, Medium) and a handful of massive tech companies. The mention rates are uniformly low, even for the most-cited sources. LinkedIn has nearly 10,000 citations but a below-baseline mention rate of 4.9%.
The takeaway for marketers: if you're evaluating your AI visibility based on top-of-funnel prompts, you're measuring noise. These prompts are almost guaranteed to return low mention rates regardless of how well-optimized your content is. Only companies with enormous brand authority - think Microsoft or IBM - can reliably appear here, and even they barely beat the baseline.
Finding 2: The Real AI Visibility Opportunity Is Mid-Funnel and Bottom-Funnel
When we move down the funnel to consideration-stage prompts ("What are the best endpoint security platforms?") and decision-stage prompts ("Compare CrowdStrike vs. Palo Alto Networks"), the picture changes dramatically.
Mid-funnel mention rates jump to 43.0%. Bottom-funnel mention rates reach 73.5%. AI goes from almost never mentioning brands to mentioning them in nearly three out of four responses.
The sources driving these mentions are fundamentally different from top-of-funnel.
Mid-Funnel: Review Sites Dominate
At the consideration stage, where buyers are evaluating options, the highest-impact sources are review and comparison platforms:
| Source | Type | Citations | Mention Rate | Lift |
|---|---|---|---|---|
| peerspot.com | Review | 7,921 | 77.4% | +34.4 pp |
| softwareadvice.com | Review | 2,179 | 79.5% | +36.5 pp |
| prnewswire.com | News | 2,178 | 86.4% | +43.4 pp |
| crn.com | News | 4,207 | 85.5% | +42.5 pp |
| trustradius.com | Review | 2,990 | 71.6% | +28.6 pp |
| gartner.com | Company | 5,411 | 71.2% | +28.2 pp |
| capterra.com | Review | 4,694 | 65.0% | +22.0 pp |
| slashdot.org | News | 8,025 | 66.8% | +23.8 pp |
| g2.com | Review | 10,290 | 61.6% | +18.6 pp |
Peerspot has the highest lift of any high-volume source at mid-funnel: when AI cites Peerspot, brand mention rates are 34.4 percentage points above the baseline. Software Advice and TrustRadius show similar patterns. These are not small sample sizes either. G2 alone has over 10,000 mid-funnel citations.
Reddit remains the most-cited source overall with 28,494 mid-funnel citations and a 50.4% mention rate (+7.4 pp lift), but the review platforms punch well above their weight in terms of lift.
Bottom-Funnel: Even Higher Stakes
At the decision stage, where buyers are making final choices, the same review platforms maintain strong lift:
| Source | Type | Citations | Mention Rate | Lift |
|---|---|---|---|---|
| sourceforge.net | Other | 353 | 94.3% | +20.8 pp |
| github.com | Company | 728 | 93.2% | +19.7 pp |
| peerspot.com | Review | 1,724 | 91.7% | +18.2 pp |
| crn.com | News | 770 | 91.9% | +18.4 pp |
| getapp.com | Review | 1,357 | 91.1% | +17.6 pp |
| g2.com | Review | 4,686 | 90.9% | +17.4 pp |
| aws.amazon.com | Company | 2,179 | 91.4% | +17.9 pp |
| prnewswire.com | News | 1,093 | 90.6% | +17.1 pp |
| softwareadvice.com | Review | 1,793 | 90.3% | +16.8 pp |
| capterra.com | Review | 2,320 | 89.5% | +16.0 pp |
At the bottom of the funnel, when AI cites G2, there is a 90.9% chance it will also mention a specific brand. Capterra sits at 89.5%. Software Advice at 90.3%. These platforms are essentially signal boosters for brand mentions in AI purchase conversations.
Finding 3: Review Sites Are the Biggest Untapped Lever for AI Visibility
This is perhaps the most actionable finding in the entire dataset.
For years, Google has been actively deprioritizing review and comparison sites in traditional search. These platforms were considered problematic because they were relatively easy to influence through paid placements, review campaigns, and SEO tactics. Google's various algorithm updates aimed to reduce their visibility in favor of more authoritative, "original" content.
AI platforms have not followed suit. In our data, review sites consistently show the highest lift across both mid-funnel and bottom-funnel stages. The AI is treating these platforms as valuable reference signals for brand-level recommendations.
Consider the practical implication: a company that runs a focused review campaign on G2 or Peerspot, encouraging customers to leave detailed reviews, could see a meaningful uplift in AI brand mentions without changing a single page on their own website. In fact, based on the lift numbers, investing in review site presence may deliver more AI visibility impact than months of technical content optimization.
The key review platforms, ranked by their combined impact across funnel stages:
| Platform | MoFu Lift | BoFu Lift | MoFu Citations | BoFu Citations |
|---|---|---|---|---|
| Peerspot | +34.4 pp | +18.2 pp | 7,921 | 1,724 |
| Software Advice | +36.5 pp | +16.8 pp | 2,179 | 1,793 |
| TrustRadius | +28.6 pp | -9.6 pp | 2,990 | 781 |
| Capterra | +22.0 pp | +16.0 pp | 4,694 | 2,320 |
| G2 | +18.6 pp | +17.4 pp | 10,290 | 4,686 |
| GetApp | +17.0 pp | +17.6 pp | 2,283 | 1,357 |
Finding 4: Wikipedia Drops Off, Reddit Stays Strong
Wikipedia is a fascinating case study in this data. At the top of the funnel, it's the fourth most-cited source with 5,408 citations and a slight positive lift of +3.8 pp. It's doing what you'd expect for educational, definitional queries.
But at mid-funnel, Wikipedia citations drop to 4,281 and the lift flips negative (-9.4 pp). At bottom-funnel, it drops further to 997 citations with -8.4 pp lift. Wikipedia becomes a negative signal for brand mentions as queries become more commercial. AI seems to recognize that Wikipedia is useful for definitions, not for product recommendations.
Reddit shows the opposite pattern. It's the number one cited source at every single funnel stage:
| Stage | Reddit Citations | Mention Rate | Lift |
|---|---|---|---|
| Top of Funnel | 12,886 | 9.5% | +1.6 pp |
| Middle of Funnel | 28,494 | 50.4% | +7.4 pp |
| Bottom of Funnel | 9,130 | 74.9% | +1.4 pp |
Reddit's 28,494 mid-funnel citations dwarf every other source. AI platforms are leaning heavily on Reddit discussions when evaluating which products to recommend. This makes sense: Reddit threads contain authentic, experience-based product comparisons that align well with what consideration-stage queries are asking for.
LinkedIn also maintains relevance across all stages (9,830 / 13,487 / 5,260 citations), though with below-baseline lift, suggesting it's widely cited but doesn't strongly correlate with brand mention decisions.
Finding 5: Market Leaders Have a Structural Advantage
An interesting pattern emerged with the largest cybersecurity vendors. Companies like Palo Alto Networks, Fortinet, and CrowdStrike appeared as highly-cited sources themselves - not just as brands being mentioned.
Palo Alto Networks' own domain was cited 5,406 times at top-of-funnel, 8,580 times at mid-funnel, and 1,123 times at bottom-funnel. Fortinet shows a similar pattern with 7,157 mid-funnel citations.
These companies have become so embedded in the AI's knowledge base that their content serves as a reference source for the entire category. Their websites, documentation, and thought leadership content effectively shape how AI understands and talks about the cybersecurity space, giving them a built-in advantage even in conversations about competitors.
In contrast, smaller vendors' domains appear far less frequently as citation sources, making them more dependent on third-party platforms to drive their AI visibility.
The GEO-SEO Feedback Loop
One broader pattern worth highlighting: the relationship between GEO (Generative Engine Optimization) and traditional SEO is becoming circular.
AI platforms use retrieval-augmented generation, pulling from search-indexed content to inform their responses. This means SEO signals - domain authority, content quality, backlink profiles - influence what AI picks up and cites. Content that ranks well in traditional search has a better chance of being pulled into AI responses.
At the same time, the growing adoption of AI search is changing how people find information. Marketers are generating massive amounts of AI-assisted content to compete in traditional search. This AI-generated content then feeds back into the data that AI platforms use for their responses.
AI impacts SEO. SEO impacts AI. The two are increasingly intertwined, and the companies that understand both systems will have an advantage over those optimizing for just one.
What This Means for Marketers
Based on 371,000+ data points, here's what we'd recommend:
1. Stop measuring AI visibility at top-of-funnel. A 7.9% baseline mention rate means you're mostly measuring noise. Focus your monitoring and optimization on mid-funnel and bottom-funnel prompts where AI actually recommends brands.
2. Invest in review platforms. G2, Peerspot, Capterra, Software Advice, and TrustRadius have outsized influence on AI brand mentions. A structured review campaign on these platforms may be the single highest-ROI activity for improving AI visibility.
3. Don't ignore Reddit. It's the most-cited source at every funnel stage. Authentic presence in relevant subreddit discussions - whether organic or through community engagement - matters for AI visibility.
4. Recognize that brand authority compounds. Market leaders benefit from a flywheel where their existing authority in AI training data reinforces their visibility. Smaller companies need to be more strategic about leveraging third-party sources.
5. Think GEO and SEO together. They're no longer separate channels. Your SEO investments feed AI responses, and AI-driven search behavior reshapes SEO competition. A unified strategy outperforms siloed approaches.
Methodology
- Sample: 21 B2B cybersecurity companies selling primarily to enterprise
- AI Platforms: ChatGPT, Gemini, Perplexity, and others
- Responses Analyzed: 371,000+
- Third-party Sources Tracked: 5,056 unique domains (brand-owned domains excluded)
- Funnel Classification: Prompts were categorized into Top of Funnel (awareness/educational), Middle of Funnel (consideration/comparison), and Bottom of Funnel (decision/purchase) based on intent
- Lift Calculation: Source mention rate minus stage baseline. Positive lift indicates the source correlates with higher-than-average brand mentions.
- Industry Note: These results are specific to enterprise cybersecurity. Different industries, B2C markets, and other verticals may show significantly different patterns.
Frequently Asked Questions
What is source influence in AI search?
Source influence measures how third-party citation sources affect whether AI platforms like ChatGPT, Gemini, and Perplexity mention specific brands. When an AI cites a particular source in its response, we track whether it also mentions the brand being monitored. Sources with high "lift" correlate with significantly higher brand mention rates compared to the baseline average.
Why do review sites have such a big impact on AI brand mentions?
Review platforms like G2, Peerspot, and Capterra contain structured, comparative product data that aligns closely with how AI answers consideration and purchase-intent queries. When someone asks AI "What are the best endpoint security platforms?", the AI draws on sources that compare products, and review sites are purpose-built for exactly that. Our data shows review sites deliver up to +36.5 percentage points of lift above baseline at the mid-funnel stage.
Should I stop optimizing for top-of-funnel AI visibility?
You should stop measuring based on top-of-funnel prompts, but that doesn't mean top-of-funnel content is useless. The 7.9% baseline mention rate at top-of-funnel means AI rarely mentions brands in educational queries. If you're benchmarking your AI visibility on these prompts, you'll consistently see low numbers that don't reflect your actual brand presence. Focus your measurement and optimization efforts on mid-funnel and bottom-funnel prompts where mention rates range from 43% to 73.5%.
How does Reddit influence AI search results?
Reddit is the single most-cited source across all three funnel stages in our dataset, with 28,494 citations at mid-funnel alone. AI platforms heavily reference Reddit discussions because they contain authentic, experience-based product opinions and comparisons. Having genuine presence in relevant subreddit discussions - through organic participation or community engagement - can positively influence how AI talks about your brand.
What is the relationship between GEO and SEO?
GEO (Generative Engine Optimization) and SEO are increasingly intertwined. AI platforms use retrieval-augmented generation that pulls from search-indexed content, so your SEO performance directly influences what AI cites. Simultaneously, AI-generated content is flooding traditional search, and AI search behavior is changing how users interact with Google. Optimizing for both channels together is more effective than treating them separately.
Do these findings apply to industries outside cybersecurity?
This specific research focused on 21 enterprise cybersecurity companies. While the general patterns - review site influence, funnel-stage differences, Reddit's dominance - are likely to hold across B2B tech, the exact numbers, top sources, and baseline mention rates will vary significantly by industry. B2C markets, SMB-focused products, and non-tech verticals may show very different source influence patterns.
How can I improve my brand's AI visibility without changing my website?
Based on our data, the highest-impact off-site activity is running structured review campaigns on platforms like G2, Peerspot, Capterra, and Software Advice. These sites show the strongest lift on AI brand mentions at both mid-funnel and bottom-funnel stages. Encouraging customers to leave detailed, authentic reviews on these platforms can meaningfully increase the likelihood that AI mentions your brand in purchase-related conversations.
Research by the Airfleet AI Visibility team. For more on how we track and optimize AI brand visibility and a free GEO audit, visit our audit page.
