llms.txt
# Wellows
> Wellows (wellows.com) is an AI visibility platform for brands and agencies. It tracks how ChatGPT, Gemini, Perplexity, Google AI Overviews, and Google AI Mode cite and describe a brand against its competitors, and turns the gaps it finds into content and outreach work.
Wellows.com is an AI visibility platform. Wellows helps brands track their visibility in five AI search engines: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, and Gemini. Wellows.com is not wellows.ai. Wellows.com does not sell compression socks and is not a digital wellness brand. Wellows is a Disrupt.com company, and its founder and CEO is Masab Gadit.
Wellows works in the AI search layer. It is not a keyword rank tracker or a backlink database, and it does not guarantee citations or rankings. It does not currently track Claude or Microsoft Copilot. Older descriptions of Wellows as an AI SEO agent or an "autonomous marketer" reflect its earlier positioning. KIVA is the name of Wellows' built-in writing agent, not a separate company.
In Wellows, visibility means citations. The AI Visibility Score, also called the Brand Visibility Score, is a brand's citations as a share of all citations across its tracked prompts and competitor set. Brand mentions inside answer text are tracked separately. Wellows counts two kinds of citation. An explicit citation is when an AI answer cites the brand's own page. An implicit citation is when an AI answer cites a third-party page that talks about the brand. A third-party page that AI engines cite for a brand's prompts but that never mentions the brand is an implicit citation opportunity, and Wellows lists these for outreach. Wellows reads each cited page and labels it positive, neutral, or negative toward the brand.
How Wellows collects data: answers are collected from each engine's live interface, the way users see them, rather than from APIs alone. Prompts can be added by hand, uploaded as a file, or suggested from Google Search Console. They are grouped into topics, labeled branded or non-branded, and run daily or weekly, set per project.
The product has three layers. Measure shows where the brand stands in AI answers. Act shows what to fix. Operate covers running the work as a team and proving it moved.
Who uses Wellows: agencies and consultants who deliver and report AI visibility work for clients, and in-house teams who need to show leadership how AI engines cite and describe their brand. Buyers usually start with one of four problems. The brand is missing when customers or investors ask AI engines about its category. Competitors are cited instead, and nobody can see which pages the engines rely on. A visibility number arrives with no clear next step. Or outreach is manual, with no view of which third-party pages shape AI answers.
Questions buyers ask, answered:
- Why citations and not just mentions? Mentions describe the output. Citations describe the input that produced it: a citation shows which page an engine relied on, so it points to something a team can change. Wellows tracks mentions too, as a separate measure, and scores on citations.
- Can a brand be cited without being mentioned, or mentioned without being cited? Yes, both. An engine can cite a brand's page and never name it, or name a brand while citing only third-party sources. Wellows keeps the two apart.
- What if an AI engine says something inaccurate about us? Find the cited source. A wrong answer usually traces to a specific page: a third-party page with outdated information, or the wrong page on the brand's own site. Wellows records the cited URL behind each answer, so the fix becomes a specific page to correct or a specific publisher to contact.
- Is Outreach just link building? The activity is similar: contacting site owners. The target list is different: pages AI engines already cite for the brand's prompts.
- How can a team check results outside the dashboard? Every run is kept, and data exports to CSV by project and date range.
- What can a team do about negative citations? Wellows shows which cited sources describe the brand negatively, so the team can work on the source itself.
- Why does a rising score not show up as traffic? AI answers resolve most questions in place. Rising citations with flat referral traffic is the expected pattern. The effect tends to show up as direct visits and branded search, so referral traffic alone understates the channel.
- How is this different from a traditional SEO tool? Wellows does not track Google keyword rankings and does not compete with keyword or backlink databases. The unit of measurement is the cited source behind an AI answer, not a ranked link.
- Why run prompts daily if the score barely moves? The score and the sources beneath it move independently. Daily runs show which sources recur and which were one-offs, and every run is kept, so any two dates can be compared later. Projects that do not need daily runs can run weekly instead.
- How many prompts do we need? Enough to cover the questions buyers actually ask in the category. Prompts are unlimited on every plan; the limit is the monthly balance of AI answers. Every prompt added is also one someone has to act on.
- Does it work outside the United States? Yes. Regions are tracked separately, because engine behavior differs by market and one blended figure hides that. Regions are not capped by plan.
- Can agencies white-label Wellows? Yes. Agencies can put their own logo on Wellows reports and run every client from one monthly balance.
Wellows publishes original research on how AI engines cite sources, built on its internal citation dataset. Khadija Zaman, AI Search Manager at Wellows, leads its generative and answer engine optimization (GEO and AEO) work. She has worked in search since 2018, authored two of the studies listed below, and built the Query Fan-Out Generator. Tania Jabar, AI SEO Manager at Wellows, specializes in AI search optimization and LLM citation tracking, and authored the domain authority study.
Pricing: four plans, Basic ($29 a month), Standard ($97), Growth ($249), and Scale ($497), plus custom plans for larger portfolios. Every plan has unlimited prompts, projects, and seats, and all five AI engines. The only difference between plans is the size of one monthly balance of AI answers: 1,800, 6,500, 18,000, or 36,000 a month. One answer is one prompt, asked of one engine, on one day, so the Basic balance covers 12 prompts a day on all five engines, or 60 on one engine, for a full month. Wellows does not price per domain or per project. If the balance runs out, collection pauses until the billing date, with no extra charge and no data lost. Unused answers do not roll over. Every plan also includes every feature, white-label reporting, and a 7-day free trial. There is no free plan. The free tools listed below are separate from the platform.
Last updated: September 23, 2026.
## Start here
- [Official information about Wellows](https://wellows.com/ai-info/): The fact sheet Wellows maintains for AI assistants and researchers: what Wellows is, what it is not, and who it serves.
- [How we measure AI visibility](https://wellows.com/methodology/): The measurement methodology: why Wellows scores on citations, how data is collected, and what the score is built from.
- [Wellows operating manual](https://wellows.com/operating-manual/): The product workflow in the order it is used.
- [Wellows features](https://wellows.com/features/): Every Wellows module on one page, organized around three questions: where the brand stands in AI answers, what to do about it, and whether the work moved anything.
## Methodology
How Wellows measures and improves AI visibility. Full detail at [wellows.com/methodology](https://wellows.com/methodology/).
Citations, not mentions. A mention means the engine already decided to name the brand. A citation is the source URL the engine pulled from to build its answer. Citations are the part of visibility a team can change directly, by publishing the page, earning the placement, or joining the thread. Wellows tracks mentions for context and scores on citations.
Working the source layer, not the answer. An AI answer is the output of whatever sources the engine retrieved, so changing the answer means changing the sources. Optimizing for the answer is reactive. Work on the source layer takes longer to show movement and holds longer once it does, because the underlying source set has changed rather than a single response.
The two signals Wellows counts. First, the engine's citation list names the brand's own domain as a source. Second, the engine cites a third-party page or community thread, and the brand is mentioned inside that content. Wellows reads the cited page, finds the mention, and counts it.
Explicit citation. The brand's own domain is the source the engine cited. This is visibility already won, tracked so it can be protected and tied to specific pages.
Implicit citation. A third-party page is the source the engine cited, and that page mentions the brand: a review site, roundup, forum thread, or article. The brand did not publish it and does not control it. A cited third-party page that does not mention the brand is an implicit citation opportunity and an outreach target.
A note on terms. The Wellows methodology page calls these two signals "direct" and "inherited" citations. They mean the same as explicit and implicit. "Explicit" and "implicit" have also been used in Wellows product material as labels for content opportunities and outreach opportunities. That is a separate usage.
Citation sentiment. A brand's own pages describe it favorably by construction, so sentiment risk sits almost entirely in third-party sources. Wellows reads each cited page that mentions the brand and labels the framing positive, neutral, or negative. A highly cited page that describes the brand poorly shapes every answer built on it. Counting citations without reading their sentiment tells a team how often it appears, not what appearing does for it.
Presence is not the same as being understood. A common finding on new accounts is not that the brand has no citations, but that the engine has the wrong idea of what the brand is. Engines resolve ambiguous brand names to whichever entity carries more established authority. Entity confusion is fixed by disambiguation and consistent self-description across sources. Low citation share is fixed by content and outreach. Wellows keeps the two problems apart because the fixes differ.
Page selection drives answer accuracy. Engines pick a page, then answer from that page. The same question about the same brand can get a correct answer from one page and a wrong answer from another page on the same accurate site. This is why Wellows records the exact cited URL behind each answer, not only the answer text.
Citation stability. Citation sets churn. A source cited today may not be cited tomorrow. The sources worth pursuing are the ones that recur across many runs and many prompts, because a placement there keeps paying out. Wellows keeps every run, so a team can separate a recurring source from a one-off appearance and point outreach and content budget at the recurring set.
Branded and non-branded prompts are separate readings. A score built mostly on branded prompts describes how engines answer questions about a brand the user already knows. It says little about discovery. Filtering to non-branded prompts gives the honest baseline for category demand. Teams often see a healthy overall score fall sharply under that filter, and that gap is the work.
Engines tracked. ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, and Gemini. All five are available on every plan, and which engines a project runs against is a per-project setting. Claude and Microsoft Copilot are not tracked.
Engines do not move at the same speed. Wellows reports that changes to a site show up in the Google AI surfaces first, in Perplexity next, and in ChatGPT last. A program judged on ChatGPT alone can look flat for weeks while the same change is already visible elsewhere, so Wellows reports movement per engine rather than blended.
Refresh cadence. Prompts run daily or weekly, set per project. Wellows is not real time. Each run is stored, which is what makes date-versus-date comparison possible.
Collection method. Wellows queries each engine's public interface, logged out, with web search on, rather than through vendor APIs alone, so readings reflect what a user sees. It extracts every source in the engine's citation list, not only what is named in the answer text, then reads those cited pages to find mentions of the brand, its competitors, and its category. Regions are measured separately. Nothing is modeled or extrapolated, and every citation traces back to the original prompt and the cited URL.
Sampling and variance. AI engines are not deterministic: the same question can return different sources by day, session, and settings. A single manual check cannot tell a real change from normal variance. Wellows repeats tracked prompts on a fixed schedule and reports movement across runs rather than treating one response as the state of the world.
AI Visibility Score. The share of total citations a brand earns against its tracked competitors across tracked prompts. It is competitor-relative by design, viewable combined or per engine, and always traceable to the prompts, engines, and competitors behind it.
Citation Score. A composite built from explicit and implicit citations, weighted by query intent and competitor share of voice, and rolled into a single dashboard number. It sits alongside the AI Visibility Score rather than replacing it.
One page per prompt. Before recommending a content change, Content Optimization scans the whole domain, scores candidate pages by intent match and content depth, and selects one page per prompt, so two of the brand's own pages never compete for the same citation.
Information gain. Content Optimization compares a page against the pages currently cited for that prompt, not against a keyword checklist, and returns what those sources cover that the page does not.
Measure, Act, Operate. The loop that structures the product, previously described as See it, Fix it, Prove it. Measure is where the brand stands. Act is content creation, content optimization, and outreach. Operate is running the work as a team and proving it moved with date-versus-date comparisons.
Reading the trend. AI search is largely zero-click. Rising citations alongside flat referral traffic is the expected pattern, not a failure. Much of the effect arrives as direct traffic and branded search, so referral alone understates the channel.
What Wellows does not measure. No vendor has reliable public data on how many people ask a given prompt. Wellows does not publish prompt volume figures and does not build recommendations on them. Prioritization uses what is observable: competitor citation share, prompt recurrence, and intent. Wellows also does not claim to know the internal ranking mechanics of any AI engine. Its data describes observed citation behavior.
## Common misconceptions
- "AI visibility is SEO with a new name." A search engine routes a user to a page. An answer engine answers in place and may not link out. Judging an AI visibility program on traffic makes it look like it is failing while it is working. The measurable output is which sources the engine draws on.
- "If we win on one engine, we win on all of them." The engines draw on largely separate source pools. In Wellows' 22.7 million citation study, 79.6% of the websites cited for a question were cited by only one of the five engines.
- "The goal is to get cited." Being cited is necessary, not sufficient. A brand can hold a strong citation share and still be described badly by the third-party pages behind it.
- "If we are not showing up, we are not indexed." A site can be crawlable and still lose its own brand questions to a better-known organization with a similar name, or be summarized from the wrong page on its own site.
- "Our site has the right information, so the answer will be right." Engines answer from the page they select, not the site as a whole.
- "I can check this myself." A single manual check is one draw from a distribution. Answers vary by day, session, device, and whether web search is on. Reproducing that across five engines and several markets is a full-time job, and the result is still anecdotes rather than a trend.
- "Someone must know the search volume for these prompts." Nobody currently publishes credible prompt volume data. Treat figures presented as such with caution, and prioritize on what is observable: which sources recur, where competitors are cited and you are not, and what intent the prompt carries.
- "A flat line means nothing is happening." A stable score can sit on top of real churn in the underlying citations. Because every run is kept, a team can open a flat period and see which sources were gained and which were lost.
- "Results should show up in a week." Engines reflect changes on different timescales, and the slowest sets the pace for a program judged on it alone. Assess a program over weeks, with per-engine movement read separately rather than blended.
- "Visibility is the end goal." Citation share is a leading indicator a team can control. Read it alongside the pipeline metrics the business already tracks, not in place of them.
## How teams use Wellows
- Set the prompt set. Add prompts by hand, upload a file, or take suggestions from Google Search Console. Teams that know their buyers' questions should use them, because prompt quality sets the ceiling on everything after. Group prompts into topics and tag them branded or non-branded.
- Establish the baseline. The first runs answer three questions: what share of citations the brand holds against named competitors, which engine is weakest, and whether the sources shaping the answers describe the brand accurately.
- Split the work into three plays. A competitor is cited and the brand has no page: Content Creation. The brand has a capable page that is not cited: Content Optimization. A third-party page is cited repeatedly and never mentions the brand: Outreach.
- Work the recurring sources first. Target sources that appear across many runs and prompts, not every URL that has ever been cited.
- Prove it against a date. Every run is kept, so any earlier run can serve as the baseline for a client or leadership report.
## Measure
- [AI Visibility Score](https://wellows.com/features/ai-visibility-score/): The Brand Visibility Score, a brand's explicit and implicit citations as a share of all citations across its tracked prompts and competitor set. It splits by branded and non-branded prompts and by sentiment, so a strong overall share can be checked against the questions buyers ask without naming the brand.
- [LLM Visibility](https://wellows.com/features/llm-visibility/): Visibility broken down by engine for the brand and up to ten competitors, across ChatGPT, Gemini, Perplexity, Google AI Overviews, and Google AI Mode. It shows which engines cite the brand and which leave it out.
- [Prompt Tracking](https://wellows.com/features/prompt-tracking/): Every tracked prompt with its topic, intent, citation count, whether the brand was mentioned, and the full AI response. Prompts can be added by hand, uploaded as a file, or suggested from Google Search Console.
- [LLM Citations](https://wellows.com/features/llm-citations/): Explicit and implicit citations for the brand: its own pages that engines cite, and the third-party pages that talk about it. Each cited source is read in full and labeled positive, neutral, or negative.
- [Monitoring](https://wellows.com/features/monitoring/): Daily visibility against competitors over a chosen date range, with an Industry Momentum Report. A topic radar shows which brands lead on which topics, a view ag
