Most teams assume the Reddit threads that matter are the big ones: thousands of upvotes, hundreds of comments, front page of a large subreddit. The data says otherwise. The median Reddit post cited by AI assistants has around 11 upvotes and is about two years old. This article explains why Google and large language models cite Reddit, what the cited threads actually look like, and how to work out which threads matter in your category.
The short answer: Google and AI assistants cite Reddit because it is a huge, constantly updated archive of first hand questions and answers, and because Google and OpenAI both have content agreements with Reddit. The threads that matter for your category are not the most popular ones but the ones that answer a buyer’s question directly: mostly question and answer threads, often old, usually with modest upvotes, that rank in Google or are retrieved for the related searches AI assistants run. Find them by checking which threads rank and get cited for your buying questions, then score them by intent and brand mentions.
Why do Google and AI assistants cite Reddit so often?
Google and AI assistants cite Reddit so often because Reddit offers first hand answers to specific questions at enormous scale, and because the companies behind the biggest AI search products have paid for access to it.
Reddit has the kind of content AI answers need
AI assistants answer questions by retrieving passages that address a specific need. Reddit is built from people asking specific questions and others answering from experience: “which CRM did you switch to and why”, “is this clinic worth the price”. Peec AI attributes Reddit’s prominence to exactly this: AI engines treat authentic user discussion as more trustworthy than marketing content (Peec AI). Google signalled the same direction in May 2023 when it said it would show more perspectives from people with first hand experience, including forum discussions (Google).
Google and OpenAI have content agreements with Reddit
Google agreed in February 2024 to use Reddit content to train its AI models, in a deal Reuters reported at about $60 million a year (Nieman Lab). OpenAI signed a content agreement with Reddit the same year (SiliconANGLE). In July 2024 Reddit restricted other search engines, including Bing and DuckDuckGo, from indexing new posts, while Google was unaffected (SiliconANGLE). Access shapes citations: systems that can read Reddit freely cite it more.
AI assistants search Reddit on purpose
AI assistants do not just stumble on Reddit; some target it. Peec AI found that Reddit accounts for 71% of the non branded domains ChatGPT targets with the site: search operator, and that Grok uses site:reddit.com searches in 10.5% of its chats (Peec AI). When an assistant wants real user opinions, it goes to Reddit deliberately.
How much of AI search comes from Reddit?
Reddit is consistently among the top three cited domains on every major AI search engine, typically accounting for somewhere between 4% and 13% of all citations depending on the engine.
| AI engine | Reddit’s rank among cited domains | Reddit’s share of citations | Source |
|---|---|---|---|
| ChatGPT | #2, behind Wikipedia | About 12% to 13% | Similarweb; Semrush |
| Google AI Mode | #2 or #3 | About 9% | Peec AI; Semrush |
| Google AI Overviews | #2, behind YouTube | Not reported | Peec AI |
| Gemini | #1 | Not reported | Peec AI |
| Perplexity | #1 | About 4% (citations are spread across many domains) | Peec AI; Semrush |
Rankings differ slightly between studies because each uses different prompts, dates and methods, but every study puts Reddit near the top. A 4% share can still mean first place on an engine that spreads its citations across many sources.
Is a Reddit thread that ranks in Google the same as one AI assistants cite?
Not always. A thread that ranks in Google for your keyword is likely to be read by searchers, but AI assistants often retrieve threads for related questions they generate themselves, so the cited thread may not rank for the original query at all.
AI search systems rewrite a question into several related searches, a process Google calls query fan out (Google Search Central). Ahrefs found that only about 12% of URLs cited by AI assistants ranked in Google’s top 10 for the original prompt (Ahrefs). It also found that AI Mode and AI Overviews cited Reddit at a similar rate, even though they shared only a small fraction of specific URLs (Ahrefs).
| Threads that rank in Google | Threads cited by AI assistants | |
|---|---|---|
| Found by | Searching your keywords in Google | Asking AI assistants your buying questions, several times |
| Who reads them | Searchers who click through | AI systems, which summarise them for buyers who may never visit Reddit |
| Typical match | Matches the keyword closely | Matches a subtopic of the question |
| How to track | Rank tracking filtered to reddit.com | Citation logging across engines and runs |
You need both lists. Threads that appear on both are the highest priority.
What do AI cited Reddit threads have in common?
AI cited Reddit threads are usually question and answer posts with modest engagement, a few years old, and short enough to quote. Popularity is not what gets them cited; relevance to a specific question is.
| Characteristic | Semrush (248,000 cited URLs) | Limy (200 hand coded cited posts) |
|---|---|---|
| Median upvotes | 5 to 8; about 80% have fewer than 20 | 11; about 40% have fewer than 10, and 6% have zero |
| Median comments | 11 to 19; about 70% have fewer than 20 | 33; 80% have more comments than upvotes |
| Age | Average of about 900 days | Median of 2 years; 11% older than 5 years |
| Format | Q and A threads are over half; with comparisons and discussions, nearly three quarters | 79% question and answer |
| Length | Median post body of about 80 words | About 2.5% had no usable body text; answers came from title and comments |
Sources: Semrush, October 2025 data across ChatGPT Search, AI Mode and Perplexity; Limy, cited posts from August 2025 to July 2026.
The practical reading: the answer usually lives in the comments, not the original post. A thread with a short question and a dozen detailed replies is a typical citation source.
Do upvotes matter for Reddit threads to be cited by AI?
Upvotes appear to matter very little. In both studies, most cited posts had fewer than 20 upvotes, and Limy concluded that upvotes are not treated as a retrieval signal (Limy; Semrush).
This changes how you should prioritise. A quiet thread in a niche subreddit that answers “best inventory software for a small pharmacy” can matter more to your pipeline than a viral post in a large general subreddit. It also means buying upvotes is not just against Reddit’s rules; it is pointless for AI visibility.
How old are the Reddit threads AI assistants cite?
Cited Reddit threads are often years old: around 900 days on average in Semrush’s data and two years at the median in Limy’s, with some more than a decade old (Semrush; Limy).
Old threads keep working because they have ranked for years and still answer the question. The risk is that they contain outdated facts: old prices, discontinued plans, problems you fixed long ago. When an AI assistant repeats an old complaint about your product, the source is often a thread from years ago that nobody has updated. Adding a current, disclosed, factual reply to such a thread is one of the highest value Reddit actions available.
Which subreddits matter for AI citations?
Subreddit size matters less than topic fit. Limy’s cited posts came from more than 100 subreddits ranging from about 65 to 3.8 million weekly visitors, including several with fewer than 10,000 weekly visitors (Limy).
- Category subreddits where your product type is discussed, such as software, finance or home improvement communities.
- Role or industry subreddits where your buyers talk shop, such as accounting, real estate or nursing communities.
- Local and expat subreddits for location based services, such as city or country communities.
- Brand and competitor subreddits, where they exist.
Large general subreddits rarely produce the specific answers AI assistants retrieve for commercial questions. Focus on where your buyers ask detailed questions.
Which types of Reddit threads matter for your category?
The Reddit threads that matter most are the ones that map to the stages of your buyer’s decision: problem, options, comparison, validation and experience.
| Buying stage | Thread pattern | Example title | What it shapes in AI answers |
|---|---|---|---|
| Problem | How do I solve X? | “How do you track inventory across two shops?” | Which solution types are suggested |
| Options | Best X for Y | “Best POS system for a small cafe?” | Which brands make the shortlist |
| Comparison | A vs B | “Tool A or Tool B for a 10 person team?” | How brands are positioned against each other |
| Validation | Is X legit or worth it? | “Is Company X worth the price?” | Trust and objections |
| Experience | Review after using X | “One year with Tool A, honest thoughts” | Sentiment and specific pros and cons |
For most businesses, options and comparison threads have the most direct effect on revenue, because they decide who is recommended. Validation threads matter late in the journey, when a buyer checks you before signing.
How do you find which Reddit threads matter in your category?
Find the Reddit threads that matter by testing your buying questions in Google and AI assistants, logging every Reddit thread that appears, and scoring each one for citation, ranking, intent, brand mentions and activity.
- List 20 to 30 buying questions across the stages above, using sales calls, support tickets and People Also Ask.
- Search each in Google and note Reddit results, including the Discussions and forums section. Use site:reddit.com searches for extra coverage.
- Ask each in ChatGPT with search, Gemini, Perplexity, AI Overviews and AI Mode, three times each, and record every Reddit URL cited.
- Search Reddit and Google for your brand and top competitors to catch threads that mention you.
- Score every thread with the priority score below and sort the list.
| Signal | Points | How to judge it |
|---|---|---|
| Cited by an AI assistant | 3 | Appears in at least one AI answer; strongest if seen across runs or engines |
| Ranks in Google’s top 10 for a buying query | 2 | Check positions in Google for your buying questions |
| Buying intent | 2 | Title or question includes best, vs, alternative, review, worth it or legit |
| Mentions you or a competitor | 1 | Any brand named in the post or comments |
| Still active | 1 | New comments in the last 12 months |
Threads scoring 7 or more out of 9 deserve action now: a helpful, disclosed contribution or a factual correction. Threads scoring 4 to 6 should be monitored monthly. Below 4, ignore them for now. For the wider method of mapping all sources, not only Reddit, see our guide to AI citation source mapping.
Do English Reddit threads affect AI answers in other languages?
Yes. Peec AI found that on Google’s AI surfaces in non English markets, between 40% and 73% of Reddit citations pointed to Reddit’s machine translated pages (Peec AI). An English thread about your product can therefore be cited in an answer given in another language.
For businesses in the UAE, India and other multilingual markets, this means English Reddit threads can shape answers in Arabic or Hindi too. Reddit’s international audience is also growing quickly: international daily users rose 28% year over year to 77.1 million in Q2 2026 (Reddit). Test your buying questions in each language your customers use and check which threads are cited.
What should you do with the Reddit threads that matter?
For each high priority thread, decide whether to contribute, correct or monitor, and act through a disclosed account that follows the subreddit’s rules.
| Situation | Action |
|---|---|
| Thread asks a question you can answer well and you are not mentioned | Add a detailed, genuinely helpful answer; mention your product only if relevant and disclose your affiliation |
| Thread contains outdated or wrong facts about you | Reply politely with the correct fact and a source, from a disclosed account |
| Thread contains a valid complaint | Acknowledge it, explain what changed, and fix the cause |
| Thread is positive and accurate | Leave it alone; do not pile on |
| Subreddit bans vendor participation | Do not post; improve your own pages and other sources instead |
The full participation playbook, including accounts, posting rules and measurement, is in our hub guide to Reddit for SEO and AEO. To check whether your work changes AI answers, track your mentions per 100 queries before and after, and see how retrieval works in how ChatGPT chooses sources.
How Crawl To Click maps the Reddit threads that matter
Crawl To Click is an SEO and AI visibility consultancy. When Reddit influences a client’s category, we build the thread list described here: buying questions tested across Google and the main AI assistants, every cited and ranking thread logged, and each one scored for priority. Clients get a short list of threads to contribute to or correct, the subreddits worth following, and a baseline of how often and how accurately they are mentioned, so the effect of any Reddit work can be measured. The same method applies to a software company, a clinic, a property developer or an online retailer; only the questions and communities change. Learn more about our approach.
FAQs on why Google and LLMs cite Reddit
Why does ChatGPT cite Reddit so much?
ChatGPT cites Reddit because Reddit contains first hand answers to specific questions, OpenAI has a content agreement with Reddit, and ChatGPT sometimes targets Reddit directly with site searches. Reddit is the second most cited domain in ChatGPT after Wikipedia.
Do Reddit upvotes affect whether AI assistants cite a thread?
Upvotes appear to have little effect. Studies by Semrush and Limy found most AI cited Reddit posts had fewer than 20 upvotes, with a median between 5 and 11, so relevance to the question matters more than popularity.
How old are the Reddit threads that AI tools cite?
AI cited Reddit threads are often old: about 900 days on average in Semrush’s data and two years at the median in Limy’s, with some cited threads more than five years old.
What kind of Reddit posts get cited by AI?
Question and answer threads make up most AI citations from Reddit, followed by comparison and discussion threads. The useful answer is often in the comments rather than the original post.
Does subreddit size matter for AI citations?
Subreddit size matters less than topic fit. Cited posts come from subreddits of all sizes, including small communities with fewer than 10,000 weekly visitors, as long as the thread answers a specific question.
How do I find which Reddit threads matter for my business?
Test your buying questions in Google and in AI assistants such as ChatGPT, Gemini, Perplexity and Google’s AI features, log every Reddit thread that ranks or is cited, add threads that mention you or competitors, and score each for citation, ranking, buying intent, brand mentions and recent activity.
Want to know which Reddit threads shape AI answers about your brand?
We will test your buying questions across Google and the main AI assistants, map every Reddit thread that ranks or gets cited, and give you a scored list of what to act on. Request a Reddit thread audit.
Disclaimer: Statistics are reported as published by the linked sources for the dates stated; studies use different prompts and methods. Reddit, Google and AI providers change their systems often. The priority score is a practical framework, not a published standard. Examples are illustrative.
