# We Analyzed 107 Million AI Answers. Here’s Where Each Industry’s Citations Actually Come From > Source: > Published: 2026-08-27 13:37:27+00:00 *This post was sponsored by Trendos. The opinions expressed in this article are the sponsor’s own.* Every AI answer is assembled from a small set of sources the engine leans on for that question. Which sources make the cut is not fixed. It shifts from one industry to the next. So if you want AI to recommend your brand, the place to start is knowing which sources it pulls from in your market. We looked at 107 million AI answers this year and pulled the top 10 citation sources for three industries (IT & solutions services, consumer goods, and communication services) across the four engines search marketers watch most: ChatGPT, Perplexity, Gemini, and Google AI Overviews. Then we sorted every cited domain into three groups: community and user generated content (Reddit, YouTube, forums), brand and retail (Amazon, vendor sites, product pages), and independent editorial and reference (media, review directories, Wikipedia). That gives a clear picture of where your effort pays off. It depends on what you sell, and plenty of brands are working on the wrong group of sources. ## Current Breakdown Of AI Citation Sources, By Industry Community and user generated content makes up almost exactly half of the leading citations in all three industries. Not high in one category and low in another, but close to 50% across the board. Industry | Community & UGC | Brand, retail & owned | Independent editorial & reference | IT & solutions services | 51% | 2% | 47% | Consumer goods | 50% | 46% | 4% | Communication services | 50% | 10% | 40% | Figures show the average share of each engine’s top 10 citation sources across ChatGPT, Perplexity, Gemini, and Google AI Overviews, weighted equally across engines. Rows are rounded and may not total 100. ### What Are Community & UGC Sources? Community content & user generated content (UGC) is content that is created by people in public, and typically, content is created for: - Reddit. - YouTube. - Quora. - Forums. - Social media platforms. ### What Are Brand, Retail & Owned Sources? Brand, retail & owned sources refer to commercial surfaces that a company or seller controls. These include first party sites and marketplace listings, such as Amazon, price comparison resources. ### What Are Independent Editorial & Reference Sources? Independent editorial & reference sources are third parties that have no stake in the sale. These are sources that span press, Wikipedia, and B2B directories. Reddit, YouTube, and forums sit under every category. Whether you sell enterprise software or car wax, public conversation between real people makes up about half of the leading citations. If your brand is absent from those threads, you are missing from half the sources AI reaches for. The interesting half is the other one, where the three industries stop looking alike. ## B2B Software & IT Services: Get Cited On Review Directories, Not Your Own Site In IT and solutions services, your homepage will not carry you; third party proof will. Getting a feel for how [AI engines decide what to cite](https://trendos.com/?utm_medium=publisher&utm_source=article&utm_campaign=sej-article-08-25) tends to separate the brands that get recommended from the ones that get skipped. Brand owned sources, such as your web pages, all but disappear at 2%, the lowest of the three industries by far, while independent editorial and reference climbs to 47%. Much of that is B2B review directories: Clutch, G2, GoodFirms, Gartner, SourceForge, and Slashdot. Ask AI to recommend a software vendor and it does not take the vendor’s word for it. It goes looking for outside validation. That rewards very different work than consumer goods does. ### Claim & Complete Your G2, Clutch & Gartner Profiles Claim and complete your profiles on the directories that matter in your niche, starting with G2, Clutch, and Gartner. A quick way to start: **Claim your listing** on the two or three directories your buyers actually shortlist from, based on your category rather than the biggest name.**Fill every field**: categories, descriptions, features, integrations, pricing, and screenshots. Half empty profiles get skipped by buyers and engines alike.**Match your category** and feature wording to the terms prospects use when they ask AI, so the listing reads as an answer to their question. ### Earn Current Reviews On The Directories AI Quotes Most of the work here sits on other people’s platforms: Earn current reviews on those profiles, because a strong directory listing is a source AI quotes directly. To keep reviews working for you: **Ask** recent, satisfied customers for reviews on the specific directories that surface in AI answers, not only the ones you prefer. **Aim** for recency and steady volume together. A regular trickle of current reviews beats one big batch that then goes stale. **Reply** to reviews, including the critical ones. Your responses add fresh, quotable text to the page. ### Pursue Trade Media Coverage In Your Category Pursue earned coverage in trade media (the PCMags and TechRadars of your category), which sit right alongside the directories in the citation lists. Where to focus: **List the trade** and consumer outlets that already rank for your category, then work that short list rather than chasing everyone.**Pitch something citable**: original data, a genuine product comparison, or an expert take, not a press release.** Make sure the piece names your product clearly**, since that exact name is the string an engine repeats. ## Telecom & Media: Wikipedia And Press Coverage Feed The Answer Communication services leans public. Community and user generated content leads at 51%, independent editorial and reference follows at 40%, and brand and retail is lowest at 10%. Reddit is the top source in three of the four engines here, and Wikipedia leads one engine’s entire list. Media names recur throughout, alongside entertainment reference sites. This is a category where public reputation, press coverage, and a solid encyclopedic presence feed the answer more than any page you publish yourself. For a telecom or media brand, almost everything that matters is earned, not published: - Maintain an accurate, well sourced Wikipedia entry, since it is quoted directly by several engines. - Invest in trade and consumer press coverage rather than another landing page. - Watch and improve how your brand is discussed publicly, because that discussion becomes the citation. ### Maintain An Accurate, Well-Sourced Wikipedia Entry Getting your Wikipedia entry right: **Check** whether your brand has an entry and whether the key facts are accurate and current. - Where facts are wrong or missing, **propose changes** backed by independent, reliable sources and follow Wikipedia’s editing rules. Do not edit your own page directly, since Wikipedia’s conflict – of – interest guidelines discourage it. **Keep the important facts**(ownership, coverage, services, milestones) sourced to third parties. Unsourced claims get removed, and anything removed cannot be quoted. ### Invest In Trade And Consumer Press, Not Another Landing Page Where the budget goes further: **Redirect** some content spend from yet another landing page into earned coverage in outlets AI already cites. **Give** journalists a reason to name you: proprietary data, a real customer story, or commentary on a live industry issue. **Track** which outlets get quoted in AI answers for your category, then prioritize those. ### Monitor How Your Brand Is Discussed Publicly Keep an eye on the conversation by: **Watching** the public discussion that feeds answers: relevant subreddits, forums, and review threads. **Fixing** the fixable. Answer questions, correct errors, and support customers where they are already talking. **Feeding** what you learn back into your press and product work, since that discussion becomes the citation. ## Ecommerce & Consumer Goods: Your Product & Marketplace Pages Are The Citation Consumer goods is the outlier. Once you set community aside, brand, retail, and owned sources take almost all of what remains (46% of the leading citations), while independent editorial nearly vanishes at 4%. There is very little room for third party media in the answer. The engines split on this. Gemini’s consumer goods list is led by Amazon storefronts and specialist vendor sites, while Google AI Overviews puts YouTube and Reddit on top for the same industry. So the engine, as much as the product, decides whether your listing or your reviews get quoted. If you sell physical products, the work is mostly on your own surfaces. ### Audit Your Amazon And Marketplace Listings First Audit your marketplace listings first. Titles, bullet points, specs, and reviews are the citation for several engines, so treat them as answer copy, not only as product copy. A quick audit to use: **Read** your top listings the way an engine does: title, bullets, specs, Q&A, and reviews, in that order. **Make** sure the details shoppers ask AI about (size, materials, compatibility, use cases) are actually stated on the page, not left implied. **Write** the listing as answer copy: clear, specific, and complete, rather than keyword-stuffed. ### Fill The Review Gap On Your Product Pages Fill the review gap. Structured, current reviews on your product pages and on Amazon feed the answer more than a blog post will. To close it the gap: **Add structured, current reviews** to your own product pages and keep your marketplace reviews fresh. **Prompt real buyers for reviews** that mention specific use cases, since those details are what AI pulls into an answer. **Surface ratings and review counts** on the page so they are easy for engines to read. ### Build The Second Front On YouTube & Reddit Build the second front on YouTube and Reddit, where the rest of the answer comes from. In consumer goods, your commercial surfaces carry the load. Get them clean and complete before you write another blog post. Where to put your effort: **Get** your products into honest review videos and into the subreddits where your category is discussed. **Support** creators and communities with accurate information and product access rather than scripted praise, which the platforms and AI both discount. **Track** what people ask and close the gaps in your listings based on the questions that keep coming up. ## How AI Overviews, ChatGPT, Perplexity, And Gemini Choose Sources Averages hide how differently each engine behaves. The makeup of a top 10 depends as much on the engine as on the industry. Engine | How it cites | Google AI Overviews | The most social. YouTube and Reddit dominate every industry, often four fifths of the top 10. | Perplexity | Community plus proof. Reddit and YouTube up top, review directories and media close behind. | ChatGPT | The most editorial. Wikipedia, trade media, and reference lead, with Reddit high on the list. | Gemini | The most industry dependent. Editorial and reference for services, but leans to brand and retail sites for physical products. | “Optimize for AI search” is several jobs, not one. Ranking in Google AI Overviews comes down to YouTube and Reddit. For retail heavy results, it is your product and marketplace pages. For the more editorial engines, it is Wikipedia and the trade press. Same brand, different playbook each time. ## Reddit Is The One Source Cited Across Every Industry Across all four engines, Reddit lands in the top three citation sources for every one of the three industries, and it is the single most cited source in more than half of those cases. No other domain comes close to that consistency. If you change one thing after reading this, make it Reddit. Look at how your brand shows up in the subreddits that matter to your market. Spamming them backfires fast, so the goal is to understand what people already say about you in the threads AI is quoting. ## How To Audit Where AI Search Cites Your Industry Industry averages point you in the right direction, but your real picture depends on your prompts, your competitors, and the engines your buyers use. You can map it yourself: - Write down the real questions your customers ask AI when they are close to choosing. - Run each question in every engine that matters to you. - Record the sources cited in each answer, one row per domain. - Tally which domains recur, and sort them into community, brand and retail, and editorial and reference. - Mark where your brand appears, where a competitor appears and you do not, and which group is driving the answer. Do that across four engines and a few hundred prompts and you have a genuine citation audit. It is also slow, and it goes stale the moment the engines update. That is the reason we built Trendos, and if you want [an easier way to see which sources each engine cites for your market, ](https://trendos.com/?utm_medium=publisher&utm_source=article&utm_campaign=sej-article-08-25)it does this tracking continuously so you can act on it instead of rebuilding the spreadsheet every month. The brands that win in AI search are not guessing at what the engines read. They watch it engine by engine, question by question, and go to work on the sources before the answer gets written. Image Credits Featured Image: Image by Trendos. Used with permission.