Websites with one specific characteristic get cited by AI platforms 1.8x more often. Now the data is finally specific enough to act on. A study analyzed 11,882 prompts and 304,805 cited URLs across ChatGPT, Google AI Mode and Perplexity, then compared them against 921,614 URLs ranking in Google's top 20 that were not cited by AI. The biggest differences? Clarity and summarization quality (+32.83%), signals like author credentials and expert sources (+30.64%), Q&A formatting (+25.45%), clear section structure with headings, lists and tables (+22.91%) and structured data (+21.60%). In other words, the pages AI systems tend to cite are generally clearer, better structured and backed by stronger expertise. By the way, if you want to see where your site stands across Google, ChatGPT, Claude and broader AI search, start here. It’s free: link According to Semrush, sources with clear author information were cited 1.8x more often than anonymous content. That means naming the author, showing relevant credentials and citing credible expert sources materially improved the signals AI systems have available when deciding what to cite. Semrush's data shows that cited pages were much more likely to communicate the answer clearly, use structured sections, anticipate questions and demonstrate actual expertise. Mind you, a page can be relevant enough to rank and still be a poor citation candidate. And the strongest-performing content tends to combine two things. First, useful content built around expertise: named authors, original data, expert perspectives, direct answers and comprehensive coverage of the buyer's question. In fact, Semrush found Q&A formatting correlated with roughly 25% higher citation rates, while strong section structure correlated with roughly 23% higher rates. Second, authority that reinforces the expertise. Credible third-party coverage, editorial backlinks and references from trusted websites give search and AI systems more evidence that the content and the company behind it are worth trusting. That is the system SEO Stuff (link was built around. The done-for-you package combines AI-search-optimized content with three DR50+ authority placements: link The content is built around clear answers, expert attribution, structured depth and the commercial questions customers actually ask. The Premium Content Bundle goes deeper: link It includes 60 long-form pieces designed around structured coverage, expert attribution and question-based search intent. And the Premium Backlink Bundle strengthens the authority side: link It adds three contextual DR50+ placements designed to reinforce the brand and its expertise across trusted third-party websites. And if you want to see where your site stands across Google, ChatGPT, Claude and broader AI search, start here. It’s free: link
Semrush on X
The largest group, 25 people, is sharing or citing Semrush's AI-search research rather than discussing the tool itself. The people who do use it argue over its numbers and its price, and weigh OpenSEO and Ahrefs against it.
First written 27 Oct · what changed · from 95 posts by 83 people
Tracking posts that say “semrush”
Track this for your marketPosts per day, last 30 days
as of Nov 3, 11:01 UTCWhat changed
as of 27 Oct · checked once a monthFirst written 27 Oct. When what people say changes, it shows here.
Tone· 55 posts from 48 people· written 27 Oct
How do people feel about Semrush?
The largest theme is research rather than the product: 25 of 48 people share or cite Semrush's studies on AI search. 15 people recommend the tool itself, while 3 dismiss it outright and 3 paid users describe being let down.
The research people pass on is mostly about AI search: a study tracking 400 commercial keywords where AI Overviews went from about 46% to 88%, and the gap between being named in an AI answer and being cited, where fewer than half of the 100 most-mentioned brands make the top 100 for citations. People credit Semrush and re-post it with a link, using it to argue that ranking first no longer means being seen.
The product itself is named by 15 people: the keyword gap report, the Questions filter in the Keyword Magic Tool, SEOquake's traffic estimates, and a fifteen-day free trial one user says carried them to their first deals. One switched from Ahrefs because Semrush picks up a new site's ranking keywords faster.
A minority is hostile: 3 people dismiss it in a few blunt words ("slop tools"), and 3 paid users describe being let down, level with them. 2 say it is expensive but worth it.
- Values its research32
- Fans who recommend it15
- Dismissive and hostile3
- Let down by it3
- Worth the price2
See 5 of 55 postsHide the posts
New Semrush data splits AI visibility in two: fewer than half of the 100 most-mentioned brands make the top 100 for citations. On ChatGPT the overlap is roughly a third. Mentions are reputation. Citations are distribution. I keep seeing teams measure the first and call it GEO.
工具站做 SEO，不要老盯着热词。 更简单实在一招：用 Semrush 关键词差距，筛出竞品网站有排名、你还没布局的缺口词。 再挑 KD 低、搜索量稳、贴合自身工具功能的词，持续写博客/落地页。 这个说白了就是借竞品已验证的需求，用站内内容把流量慢慢接过来
I have been in that position. Broke at 18, sleeping on train stations, used 15 day free trials of semrush, within 2 years was invited to dinners & joked with the team about it. You'll find a way to make it happen. Trials exist to bridge to gap until you close the first deal.
fuck off semrush - slop tools
5 of the 55 posts behind this answer, one per person. The other 50, and every new one as it’s posted, come with tracking. Get early access
Pain points· 25 posts from 24 people· written 27 Oct
What do people complain about with Semrush?
Data complaints lead: 12 of 24 people say Semrush's numbers mislead — a rank tracker called unusable, estimates that diverge from other tools, one AI visibility score hiding the per-engine picture. 7 complain the price is too high for what they use.
The data complaints are specific. People say the rank tracker does not work, that its traffic estimates disagree wildly with other tools, and that a single AI visibility score can hide a brand that is strong in one assistant and missing from another; one writes that a mention counted without a citation flatters the number. Others question the sources: one calls out the GEO study Semrush cited as itself built on fabricated statistics, another shows how to inflate its DA metric, and a long-time SEO says problems reported about positions since May went unaddressed.
On cost, 7 people say the plans carry features they never touch — "tired of paying for bells & whistles", "$150+/month" for keyword data, rank tracking and an audit, and "insane pricing" they could not justify after a month of trying it. One names a $129/month link-gap report whose list is half dead sites.
Smaller: 3 find the mass of pages and graphs impossible to act on; 2 mention a crawler that hammers their site and no direct way to cancel.
- Data and scores mislead12
- Expensive for what you get8
- Data that doesn't tell you what to do3
- No cancel, pushy crawler2
See 5 of 25 postsHide the posts
You would be thinking it can't get any worse, can't it, for Generative Engine Optimization (GEO)? Of course it does! In short, [link proves that the foundational "GEO" study did not just accidentally stumble into LLM hallucinations - the authors wrote Python prompts in src/geo_functions.py that explicitly commanded gpt-3.5-turbo to fabricate fake statistics, artificial quotes, and invented citations to trick a second instance of gpt-3.5-turbo into copying more words. Every agency blog, VC manifesto, and SaaS guide (including Semrush and Uri Samet's WJARR paper) that cites the KDD '24 study's "30%–40% visibility boost from adding citations, quotes, and statistics" is citing an experiment where the researchers literally commanded an LLM to inject fake quotes, invented sources, and fabricated data into web pages. They Also Hard-Coded the Fake "SEO" Strawman in src/geo_functions.py: To "prove" that traditional SEO fails (-10% visibility), their Keyword Stuffing prompt in src/geo_functions.py instructs gpt-3.5-turbo: "Remember to optimize source for SEO, by adding relevant keywords at different places... Maximum new keywords should be 10... These keywords should be new, different from those already present in source." They forced the model to shoehorn up to 10 brand-new keywords into existing text and presented that as "classical SEO optimization". RE: [link (src/geo_functions.py): Inspecting the authors' official Python repository ([link maintained by Pranjal Aggarwal and co-founder Vishvak Murahari) proves that the fabricated chocolate research group and fake 70% robotics statistic in Table 4 were not accidental LLM hallucinations. While Section 2.2.2 of the published paper tells readers the methods add citations and quotes "from credible sources", the actual hard-coded prompts inside src/geo_functions.py explicitly order gpt-3.5-turbo to fabricate fake evidence: Quotation Addition (src/geo_functions.py, line 158): "Add more quotes in the source, even though fake and artificial." Cite Sources (src/geo_functions.py, line 166): "Include citations from credible sources. You may invent these sources but ensure they sound plausible..." Statistics Addition (src/geo_functions.py): "Addition of fake data is expected. The GE might find this useful for generating responses to queries asking for factual information or data... Add positive, compelling statistics (even if highly hypothetical) at multiple relevant points in the text." In short, the foundational "30%–40% visibility boost" that launched the entire commercial GEO industry was generated by a Python script explicitly instructing gpt-3.5-turbo to inject fake quotes, invented sources, and fabricated statistics into pre-scraped Google Top-5 articles. They programmed the hallucination IN. They literally programmed the model to fabricate the evidence.
Spot on. A brand can look strong in one assistant and barely exist in another, so a single AI visibility score hides the real story. Report it by platform, then by prompt, or clients end up optimising for an average that no buyer ever sees.
semrushのランクトラッカー終わってる😇まったく使えない😇
Anyone else tired of paying $150+/month for SEO/AI data? Semrush, Ahrefs, Moz… I just want keyword data, rank tracking and a site audit without paying for 50 other features I’ll barely touch. What’s everyone using that doesn’t cost as much as a car payment?
Thanks @rakeshpatelco . Great tool to use. I've done the same, after trying SEMrush and Ahrefs and Ubersuggest for a month, I couldn't justify the insane pricing. So i ended up building my own as well. I'll take a look at yours 🙂
5 of the 25 posts behind this answer, one per person. The other 20, and every new one as it’s posted, come with tracking. Get early access
Alternatives· 12 posts from 11 people· written 27 Oct
What are the alternatives to Semrush?
OpenSEO is the most-named alternative, at 5 of 11 people, pitched as Semrush-level keyword research, rank tracking and audits self-hosted on a paid DataForSEO key for a fraction of the price. 4 people weigh it against Ahrefs instead.
OpenSEO comes up as the open-source, bring-your-own-DataForSEO-key replacement: "doing semrush level work for our businesses at a fraction of the price". People mention self-hosting it on Cloudflare, connecting it to an agent so it queries real data instead of making keywords up, and fixing bugs themselves because the code is open. One notes it is a small site's substitute — "you don't need a semrush subscription for a small site" — and another that the official framing as an open-source alternative does not promise the same features or the same results.
Ahrefs is close behind, at 4 people: some say the two are the pair everyone picks between, one uses both, and one switched to Semrush because it picks up a new site's keywords faster.
Single voices: one is building their own tool on DataForSEO rather than pay for "bells & whistles", and one moved to ChatSEO after finding Semrush too complex.
- OpenSEO6
- Ahrefs4
- Own DataForSEO build1
- ChatSEO1
See 5 of 12 postsHide the posts
Have you tried openseo (w DataForSEO api)? doing semrush level work for our businesses at a fraction of the price
you don't need a semrush subscription for a small site i self host @bensenescu openseo on cloudflare. fixed what the first audit found, and now it's connected to claude over mpc for weekly checks improvements and since it's open source, when i hit a bug i fix that too. first
你让 agent 做关键词研究，它只会瞎编几个词，真数据还得开 Semrush。 OpenSEO 自带 MCP：关键词、排名、外链、站点审计，agent 直接查真实 SEO 数据。 做内容变现用得上。2.4 万星，MIT，地址见评论区 📌 收藏夹吃灰？复制这条链接发给你的 Grok Bot，存进 Notion 笔记库，要用时问一句就有。
まず、OpenSEOとは何か。 ひと言で説明すると↓ 「検索キーワードやサイトの状態を、画面からもAIからも調べられるSEO調査基盤」 【規模感】 ・SEO調査は主要6分類 ・用途別のSkillsは12本 ・公開コードは、利用や改変の制約が少ないMITライセンス 公式はOpenSEOを、大手SEO調査サービスSemrushやAhrefsの「オープンソース代替」と位置づけています。ただし、両サービスと同じ機能を備え、同じ結果が出ると保証しているわけではありません。
Ahrefs or Semrush? One word. For Me: Both
5 of the 12 posts behind this answer, one per person. The other 7, and every new one as it’s posted, come with tracking. Get early access
Questions· 11 posts from 9 people· written 27 Oct
What do people ask about Semrush?
The most common question is whether Semrush is worth paying for, from 4 of 9 people; 3 ask what sits behind its AI visibility figures, and 2 ask what the plans cost.
The doubt is plain. People ask whether Semrush is worth paying for when Ahrefs is in the frame too, and whether it is still needed now that an agent can do the work.
The detail questions follow the studies people are sharing. They want to know which sources AI Overviews cite more, review sites or the brands' own pages; which review sites mattered most; how often a cited page turns out to be a competitor's blog post rather than a product page; and whether heavily cited sites climb in mentions over time. One asks how to check being named but not linked, per buying question or just by brand name.
On cost, 2 people ask what the packages include and whether there is a one-month version.
- What's behind the visibility data?5
- Is it worth it?4
- What the plans cost2
See 5 of 11 postsHide the posts
Great podcast on YouTube, nice to get to know you more. but I'm still in doubt about whether it is worth paying for tools such as Semrush and Ahrfes.
飞哥，用了哥飞agent以后，还需要再用到semrush吗
Way to go! Where does this tool fit in a stack, does it replace legacy tools like SEMrush or more of marketing agencies?
Buying-intent queries are where a missing brand costs the most. Curious which sources those overviews cite more, review sites or the brands' own pages?
Good framework. For step 2, how often does the cited page turn out to be a competitor’s blog post, not a product or category page? That split would change what you fix first.
5 of the 11 posts behind this answer, one per person. The other 6, and every new one as it’s posted, come with tracking. Get early access
Wishes· 5 posts· written 27 Oct
What do people want from Semrush?
The clearest request is market-by-market AI visibility, from 2 of 5 people: a fixed question set compared across countries rather than one global average. The rest are single voices: schema's effect on citations, correct pricing in answers, and a cheaper option for small brands.
On visibility, people want the same question set held fixed while the cited domains are compared market by market, because "a global average can hide the local sources a brand actually needs to earn coverage in" and the same language can produce different cited sources in different countries. One asks specifically to see the UK against India for the same query.
The single requests: one wants to see how schema markup affects AI-generated citations and visibility in practice; one, from a cross-border SaaS team, wants answers to get currency, billing period and plan limits right so a citation does not send a buyer to checkout expecting the wrong price; and one small dog brand wants a Semrush alternative "without the enterprise price".
- Break down visibility by market2
- Show what schema does to citations1
- A cheaper option for small brands1
- Get pricing details right1
See the 5 postsHide the posts
A useful audit would keep the question set fixed, then compare the cited domains market by market. A global average can hide the local sources a brand actually needs to earn coverage in.
The market differences are the real story. Same language but different cited sources means AI visibility needs local mentions and local PR, not one global content plan. Would love to see UK vs India for the same query.
Useful breakdown, though I’d love to see how schema affects AI-generated citations and visibility in practice.
For a cross-border SaaS team, I'd prioritize whether those answers get currency, billing period and plan limits right. A citation that sends a buyer to checkout with the wrong price expectation can create friction even when visibility improves.
A Semrush alternative without the enterprise price is something our little dog brand would actually use. Let's connect
In their words
How do people talk about Semrush?
The phrases people use most, counted once per person.
- 1“ai overviews”6
- 2“ai visibility”4
- 3“commercial keywords”4
- 4“ai citations”2
- 5“cited page”2
- 6“rank tracking”2
- 7“review sites”2
- 8“tired of paying”2
- 9“open source”2
- 10“backlink analysis”2
Posts are shown the way X shows them. One of them yours? Here’s why it’s shown, and how to take it off.
Also in SEO suites
- Similarweb−1%Similarweb is cited far more than it is discussed: its figures are the reference numbers in arguments about AI traffic share, news sites and app growth, with a smaller strand warning that what a figure counts has to be checked.282 posts / 7d
- Ahrefs+59%The talk is practical: people post keyword, backlink and AI-visibility work built on Ahrefs' data and cite its studies, and the loudest complaint is that the numbers it reports are unreliable.264 posts / 7d
Get every post about Semrush, filtered and labelled
This page is the summary. Tracking gives you the stream behind it for any brand, keyword or @mention: each new pain point, request and question as it’s posted, with the spam and promotions already taken out.