Use case · Ranking
When the AIs recommend, who do they name?
On the prompts your market actually asks, Epovest measures who the engines mention, how often, and how that share moves check after check.
Views from the app. Your figures replace the sample at the first survey.
Your market, read as a scoreboard
Ranking tracks several keywords: you and the names you compare yourself with. Each check asks your prompts to the engines and counts who is mentioned. You read a citation rate per keyword and per engine, and a share of voice across the whole set: not an impression, a measure.
What you measure
Keyword scoreboard
The citation rate of every keyword you track, engine by engine and over time.
Share of voice
Your share of all tracked-keyword mentions on your prompts.
Cited sources
The domains the engines lean on when they answer: where the recommendations come from.
Discoveries
A domain that keeps coming back in the answers and that you do not track yet is suggested. Nothing is added without your say-so.
The recipe
How to work your AI ranking with Epovest
The full working guide for this use: buying prompts without your name, an honest keyword landscape, and the Atlas as your battle plan for share of voice.
The prompts: your market's buying questions, not your name
The rule that defines this use: your prompts contain NO tracked name, or you whisper the answer to the engine. You ask the questions a buyer asks right before choosing.
"Which [accountant] would you recommend in [Lyon]?"
"What is the best [payroll software] for a small business?"
"What should I use instead of [the name your market cites first]?"
"An affordable [SEO consultant] for an online store: who would you pick?"
Three to seven prompts, one per buying intent: recommendation, segment, alternative to the incumbent, budget or premium. Your customers' language and city matter: an engine does not recommend the same names in French and in English. The only name allowed inside a prompt is the incumbent of an alternatives prompt, and you read its line knowing that bias.
Keywords: you, and the entire landscape
Share of voice is computed over the tracked set: an incomplete landscape flatters you and teaches you nothing.
Your name as favorite, then every name the market could cite in your place, including the ones you would rather not see. Leave discovery on: the engines will cite names you did not think of, and the tracker will suggest the domains that keep coming back; accept them in one gesture so the ranking stays honest. An accepted name joins in place: its series starts at the next check, the others continue.
The settings that fit a ranking
This use compares rates between close names: every choice below serves the readability of small gaps.
- Analysts
- Presence and Share of voice, both included: the citation rate per name and per engine, and the ranking that comes out of it. Add Sentiment if HOW each name gets cited matters in your market; it is billed per analyzed answer.
- Cadence
- Weekly for a living market, monthly for a slow B2B niche, daily during a campaign or a launch. The question to ask: how fast does your market change its mind?
- Resolution
- The most resolution-hungry of the three uses: you are comparing rates between close names, and in hd a gap of a few points drowns in the draw. Settle your prompts in hd, then move to full_hd or 4k once the list is stable: repetition is what makes an overtake readable.
Read the scoreboard: per engine, not as an average
Each engine has its own habits and its own sources: an inter-engine average erases precisely the information you can work with.
Read your line against each competitor, engine by engine: being third everywhere plays out differently from being first on two engines and absent from the other two. When a name climbs, open its answers: the passage citing it almost always names the source that tipped the scale, and that is information you will reuse at the next step.
The Atlas: the battle plan for share of voice
Winning share of voice, concretely, means being mentioned on the pages the target engine already listens to. The Atlas ranks them for you.
Take the engine where your rate is weakest and sort the Atlas by its authority: the list narrows to the sources THAT engine cites, the heaviest first. Cross it with your corroborations register: a top source where no page mentions you is your priority, a directory to be listed in, a comparison to enter, a publication to talk to. Candidates show you the reverse: the pages where the engines already see you, to record without effort.
Corroborations: hold the pages that make the ranking
In this use, comparison pages are the battleground: a name that drops out of a comparison drops out of the answers a few checks later.
Record every page where your name stands next to others, comparisons, directories, awards, with its publication date. Arm weekly monitoring on the ones that carry weight: if your line disappears from a comparison, if the passage changes, if the link drops, you know it the same week and the outreach starts at once. Fill in the request channel: the file will then know to suggest which proofs to get refreshed when your canon evolves.
Quests and Logbook: share of voice is won one errand at a time
This is the most active of the three uses: every point of share of voice is an errand that succeeded somewhere.
Turn the Atlas into a work file: one quest per source to win, named after its target, get a mention on [the top source of the target engine]. In the Logbook, date every publication obtained, every campaign, every comparison entered: the share-of-voice curve then reads errand by errand, and you know what paid off.
Worked example
Share of voice is stuck on one engine
Sort the Atlas by that engine's authority: its top ten sources are its bibliography. Two of them mention you nowhere: pose one quest per source, land an entry in one's comparison and a listing on the other, record both pages as corroborations, date the publications in the Logbook. The next checks tell you whether the engine added you to its answers; monitoring then keeps both pages in sight.
With your AI assistant
Your AI assistant runs this recipe with you
Connect Epovest to the assistant you already use, one-click connector or API key. It creates trackers, reads the results and keeps the registers with you; a payment only ever happens with your approval.
The instruction to give it
Connect to Epovest and set up a ranking tracker for [my market]: five buying prompts without my name, my competitors as keywords with discovery on, Presence and Share of voice, weekly. Then tell me where I am losing and which Atlas sources to work.
$0.10
per prompt, per engine, per check
Pricing
Whatever you configure, the unit does not change. Your total shows live in the configurator before anything runs; you start whenever you want, from a $10 top-up. No subscription: only the checks that run are debited.
Add a keyword whenever you like
Track a new keyword at any time: its series starts at the next check, next to the others. Comparing more does not reprice the collection.
The other use cases
Make AIs recommend you
Create a tracker in minutes: your prompts, your engines, your cadence. The first check runs the moment you start.
No subscription. Credits from $10 that do not expire. Pause anytime: the history stays yours.