What AI Product Recommendations Really Cost: Inside GetCited's 106,758-Citation Study

The Methodology: 3,334 Questions, 106,758 Citations

GetCited, a Berlin-based agency specializing in placing comparison articles on media sites for AI recommendation, partnered with DataPulse Research and analytics platform BuzzView to measure how AI search engines source product recommendations. The study, published in September 2026, is the most granular look yet at the commercial layer behind AI answers.

The methodology was thorough. Researchers wrote 3,334 English comparison questions of the kind people ask before buying something: 'which payroll software should I choose?' or 'what pet insurance do experts recommend?' These covered 748 product and service types across 15 sectors. BuzzView ran every question through ChatGPT, Google AI Overview, and Perplexity in April 2026, logging every URL the three tools cited.

Each cited page was then loaded in a headless Chromium browser using Playwright with full JavaScript rendering, and checked against 22 markers of commercial content across three layers: affiliate disclosures, partner-content notices, and explicit advertising labels. The full corpus also included 6,668 questions (half in German), producing 139,033 unique URLs, of which 132,335 (95.2%) loaded successfully.

The English set alone yielded 106,758 citations from 8,536 AI answers. That is a massive dataset, and while GetCited has a commercial interest in the outcome (it sells the service of getting brands into comparison articles), the methodology and data are detailed enough to take seriously.

The Headline Finding: 25.4% of Citations Are Commercially Disclosed

The central figure: 25.4% of English citations lead to a page carrying a disclosed commercial interest. Most of that is affiliate content, review and comparison pages that take a commission when a reader buys through a link. Affiliate disclosures appear on 21.4% of all English citations, partner content on 2.9%, commerce sections on 2.0%, and advertising labels on 1.8%. The layers overlap, which is why they do not add up to the total.

When counted once per unique URL instead of per citation, the share falls to 20.5%. The study acknowledges that residual false positives could pull the figure down 'by a few points,' while pages with no machine-readable disclosure push the true share higher 'by an amount the study cannot measure.'

But the per-citation rate is only part of the story. The more striking finding is that 86% of AI answers contain at least one commercially disclosed page. Most AI product recommendations are not built on neutral ground.

How Each AI Engine Differs

The three platforms behave very differently in how they source recommendations:

The ranking flips with the unit of measurement. ChatGPT has the lowest commercial share per citation, yet the highest share of answers containing at least one commercially disclosed page, because it cites about twice as many sources as its rivals (averaging 19 per answer vs. 10 for Perplexity and 9 for Google AI Overview). Perplexity sits at the other extreme: one answer in six rested mostly on commercially disclosed pages.

Ten Publishers Carry the Commercial Layer

The most actionable finding for marketers: ten publishers account for roughly a quarter of all commercially disclosed English citations:

These are mostly affiliate-funded review brands whose model is declared on every page. When AI engines cite product recommendations, they return to the same publishers again and again. If your brand is not featured on one of these sites, you are largely invisible in AI product recommendations.

Which Categories Are Most Commercially Influenced

The study also reveals which product categories are most and least commercially influenced in AI answers:

Most commercially influenced (highest share of commercial citations):

• Consumer electronics: 41.3%
• Fitness and sports: 37.5%
• Telecom, VPN and hosting: 32.8%
• Media and entertainment: 31.0%

Least commercially influenced (lowest measured share):

• Finance and insurance: 19.9%
• Software and B2B tools: 19.2%
• Legal and professional services: 17.8%
• Travel and hospitality: 16.5%

But the study warns that the low end 'is not neutral ground.' Clinics, lawyers, and installers are answered largely from provider pages and directories, which carry nothing a scanner can read as a disclosure. The absence of a disclosure label does not mean the absence of a commercial interest.

What This Means for Brands and Marketers

1. The Affiliate Layer Is the AI's Default Source

If you sell a consumer product, AI search engines are most likely citing affiliate review sites, not your brand page. Getting your product reviewed (or at least listed) on TechRadar, Forbes Advisor, PCMag, or equivalent publications is now a prerequisite for AI search visibility, not a luxury. This is uncomfortable for brands that have been trying to build direct-to-consumer relationships, but it is the reality of how AI discovery works today.

2. B2B Has Less Competition and More Opportunity

Software and B2B tools have one of the lowest commercial citation rates (19.2%). That means fewer affiliate-laden review sites dominate the AI answers, which means original content from vendors and specialized publications has a better chance of being cited. If you are a B2B brand, this is your window to build AI search authority before the affiliate ecosystem fills in.

3. Transparency Is Not Great

Only 25.4% of citations carry a disclosure, but the study itself acknowledges that many commercial pages have no machine-readable disclosure. The real share of commercially influenced AI recommendations is likely higher. For consumers, this means AI product recommendations should be approached with the same skepticism as any other recommendation. For brands, it means the playing field for AI visibility is shaped by commercial relationships that are not always visible.

4. Think in Terms of Sources, Not Keywords

Traditional SEO asks: does my page rank for this keyword? AI search visibility asks: which sources does the AI cite for this question, and how do I get my brand onto those sources? The answer is often not about optimizing your own website but about being present on the sites the AI already trusts. That means PR, product reviews, expert commentary, and industry publications. It is a different playbook from classic SEO.

The Bigger Picture: AI Search Economics

The study arrives at a moment when the economics of search are shifting fundamentally. Future plc, which owns TechRadar and Tom's Guide, reported a 24% drop in eCommerce affiliate revenue and a 15% fall in website sessions for the six months to March 31, 2026. Ziff Davis, owner of PCMag and CNET, wrote off $54.8 million on its health media division in August 2026. The very publishers that dominate AI citations are struggling with declining traffic from traditional search.

Meanwhile, a randomized field experiment with 1,065 desktop Chrome users measured a 39.8% fall in outbound organic clicks when an AI Overview appeared. AI search is both citing these publishers more and sending them less traffic. That tension will eventually reshape the affiliate model, but for now, the strategy for brands is clear: get onto the sources AI cites, because that is where discovery happens.

Read Next

• ChatGPT Is Quietly Building Its Own Search Engine: What It Means for Brands
• OpenAI DevDay 2026 Preview: Managed Agents, Always-On 'O', GPT-6 Cyber, and New Platform Plans
• The Blueprint Alliance: How Okta, AWS, and Google Cloud Plan to Secure AI Agents
• AI Visibility in 2026: Which Platforms Actually Matter for Your Brand