Buyers ask AI which brands to consider. Each engine names a shortlist. keenOly asks those questions for you, on six engines, from your market, and shows you what came back.
When someone asks ChatGPT, Gemini or Google’s AI for "the best cooling mattress in Australia", the answer names three or four brands, describes each in a sentence, and cites a handful of pages it read to decide. Nobody clicks a results page. The shortlist is the result.
keenOly measures that shortlist. It writes the questions your buyers ask, asks every engine the same questions from your country, stores every answer word for word, and reads each one: whether you are named, how you are described, who is named instead, which pages the engines trusted, and what decided a head-to-head.
Then it tells you what to do about it, in order of what would move the number, and writes the first draft of the content that would.
This is not SEO. Search ranks pages; AI answers name brands. keenOly measures whether and how you are named inside generated answers, and what the engines cite to say it.
keenOly reads your site and your market and writes the questions buyers ask AI about your category. Three kinds: discovery ("best cooling mattress"), branded ("is Harbour & Co any good"), and comparison ("Harbour & Co vs Nimbus"). You check them, change any, add your own.
Every prompt is asked on the engines you track: ChatGPT, AI Mode, AI Overview, Copilot, Gemini, Perplexity. Each answer comes from a real browser session in your market, the way a buyer would see it, not from an API.
Every answer is stored in full, with its citations, the searches the engine ran first, and any ads, shopping cards or places it showed. Then it is read: brands named, in what order, in what tone, and which of your pages it cited.
One credit is one successful answer from one engine. A run costs prompts times engines: 25 prompts on 6 engines is 150 credits. Failed answers are never charged. Runs fire daily, weekly or fortnightly at 1am in your timezone, or whenever you press Run analysis.
The free scan is ten prompts on ChatGPT, every answer readable, at our expense. The other five engines answer the same ten prompts at the same time; their answers are stored and unlock for 50 credits.
Eight items in the menu, and Oly. Each answers one question, and each has a short video.
The first page after a run. One sentence on where you stand, written from the latest answers, with a chip saying how much data it rests on. Under it, every answer as a dot: rows are your prompts (or, on one engine, your rivals), columns are the engines. A filled dot is a mention, a dashed one an answer that left you out. Tap any dot to read that answer.
The dark card at the bottom is the one thing to fix first: the highest-ranked finding from your pages and answers, with "Ask us to run it" and "Fix it myself" side by side.
Every answer, organised four ways. Prompts: each question you track with one cell per engine, named or not, won or lost, so you see which questions you never appear for. Companies cited: every brand the answers name, yours and your rivals’ and the ones you do not track yet, with their share of mentions. Sources: the pages the engines read to write their answers, bucketed as your site, a rival’s, the press, a retailer or a community, and scored for influence. Branded: what AI says when asked about you by name, with the sentiment and the caveat it attached.
Open any prompt to read every engine’s answer in full, the searches it ran first, and the pages it cited.
ChatGPT shows sponsored results beside some answers; several engines show shopping cards, place listings and media. keenOly captures them on every run: who advertised against your prompts, which products and merchants were carried, which places were suggested. For a brand that sells, this is where you see whether your products reach the card and at what price.
One record per page of your site: how often the engines cite it, on which prompts, what Google sends it (when you connect Search Console and Analytics), and what the technical audit found. The audit checks what an AI crawler meets: robots rules, structured data, the stack, whether the page answers at all. The score is a readiness score, and an inconclusive audit says so rather than passing.
Content on Autopilot reads the questions where you are missing and proposes posts that would answer them well enough to be cited. You choose, Oly writes a draft in your own voice, learned from your site, and nothing is used until you approve it. Connected to WordPress, approved posts go over as drafts or scheduled for the next day. Pages the engines already cite without naming you come first: they are trusted, and only the name is missing.
The content review reads your existing pages for copy that reads as templated or machine written, and offers a rewrite that keeps every real fact.
Two lists, ordered by what would move the number. Findings are facts about your pages and answers that a rule detected: a page that converts and is invisible to AI, a page that ranks in Google and is absent from answers, a competitor cited where you rank, a cited page that is thin, a dead page still linked, structured data missing, a lookalike domain cited for your brand. Each carries its importance, its risk, a lane (self-serve, GEO Services or Managed GEO) and the steps.
Recommended actions are the sources: your own pages the engines fetch but never quote, third-party pages where you are missing, pages you already dominate, and a rival’s pages you can only answer with your own content. Each can have a brief drafted: target prompts, the competing page’s angle, the gaps, an outline and claims flagged for verification.
A weekly or monthly report with the same numbers the pages show: the headline figures, what to do next, by engine, by prompt, run by run, the sources and the Google data. Share it by link, download it as a PDF, or on a paid plan as an executive deck. The report counts the whole week or month; the Sources page counts the latest run, and both say so.
Tracking: your brand name and its other names, the market every run asks from, the schedule, the engines. Integrations: WordPress, Google. Team & profile. Plan & billing, where credits are the only meter. Earn credits and Request GEO Services open from here too.
Oly is the assistant inside keenOly, and a mode rather than a widget: the switch at the top of every page turns the dashboard into a conversation about your own data. Ask why a number moved, which rival wins a prompt, what a source is, or what to do first, and Oly answers from your tracked runs with a link to the page. Oly can also act, and every action waits for your yes: track a competitor, draft a brief, start a post, request GEO Services.
The free scan includes ten questions. Paid plans include unlimited questions.
Twenty minutes a week is enough. The product does the asking and the reading; your job is to decide.
Every finding carries a lane. The lane is not a paywall: it is a judgement about risk, made the same way every time.
Risk is what the fix touches. A structured data block touches nothing the visitor sees (low). Copy or structure on a page that is already earning can lose what it earns (medium). URLs and redirects can lose rankings outright (high). Risk goes up one level when the page is one of your top ten by clicks on page one, or by conversions.
Low risk is self-serve, unless the finding’s importance is high, in which case a second pair of eyes is worth it. Medium risk always goes to GEO Services. High risk is Managed GEO when the importance is high, GEO Services otherwise.
Two rules are fixed: a dead page that is still linked is always self-serve (it is one redirect), and "strong in Google, absent from AI" is always Managed GEO.
Press Request GEO Services at the top of any page, or "Ask us to run it" on a finding, or ask Oly. Tell us what you want looked at; a finding fills the form in.
Someone on the team reads your results, not a brief, and replies, usually within one working day, with what they would do and what it costs. It is a separate engagement from your subscription, priced on scope.
Nothing on your site changes until you agree a plan. Nothing is charged for the reply.
Credits are the only meter. One credit is one successful answer from one engine.
| Plan | A month | Credits | Typical capacity | Posts by Oly |
|---|---|---|---|---|
| Free AI Visibility Scan | Free | 10 prompts, once | 10 prompts on ChatGPT, once | 1, then credits |
| Lite | $49 | 2,500 a month | ≈ 25 prompts across 3 engines, tracked daily | 5 a month |
| Growth | $149 | 10,000 a month | ≈ 50 prompts across all 6 engines, tracked daily | 10 a month |
| Elite | $299 | 25,000 a month | ≈ 130 prompts across all 6 engines, or several projects | 20 a month |
ChatGPT, AI Mode, AI Overview, Copilot, Gemini, Perplexity. Every answer comes from a real browser session, the way a buyer would see it, not from an API.
No. Monthly credits reset each billing cycle. Annual plans receive their credits upfront. Earned credits do not expire with a period.
No. Move to the next plan, or earn credits by linking to keenOly, posting about it or referring a business.
This is not SEO. Search ranks pages; AI answers name brands. keenOly measures whether and how you are named inside generated answers, and what the engines cite to say it.