Use case · Reputation
What do the AIs say when asked about you?
Your customers, partners and future hires ask the engines about your name. Epovest asks them the same prompts at your cadence, and measures whether they answer, and in what terms.
Views from the app. Your figures replace the sample at the first survey.
One name, kept under watch
Reputation watches your name. Every check asks your prompts and reads the answers: does the engine answer, does it stay factual, does a negative wording creep in? A steady answer with no negative tone is the baseline; any dip is a signal you can trace in the full answers.
What you measure
Answer rate
The share of answers that actually mention your name when the engines are asked about it.
Tone of the answers
Positive, neutral or negative, answer by answer, engine by engine, with the optional sentiment analyst.
The full answers
Every answer is archived and readable in the app: each figure traces back to a text you can check yourself.
Negative answer alert
The weekly email report flags any negative answer as soon as a check finds one.
The recipe
How to run your AI reputation with Epovest
The full working guide for this use: the prompts that make a real measurement, the settings that fit it, and what each tool contributes when your own name is at stake.
The prompts: what you ask the AIs when a name is at stake
A reputation measurement asks the engines about YOUR name, the way your customers do. Four families cover the essentials; replace the brackets and keep the spoken phrasing.
"What do you know about [your brand]?"
"Is [your brand] trustworthy? What do its customers say?"
"Who is behind [your brand], and how long has it existed?"
"[your brand], scam or legit?"
Three to five prompts are enough: one per family, in the language of your customers. The name stays IN the prompt, that is what defines this use; add one prompt per flagship product if the answers mix them up.
The settings that fit a reputation
Every choice below follows from one fact: you are watching one name for surprises, not racing anyone.
- Keywords
- Your exact name as favorite: it is the one the rates follow. As variants, everything the answers use to name you: acronym, trade name, domain, the founder's name for a personal brand. Two spellings are two keywords, so two series.
- Analysts
- Keyword presence for the answer rate, and the Sentiment analyst: tone is the heart of this use, and every negative mention is a signal. It is billed per analyzed answer. Share of voice only earns its place if you also track competitors, which is Ranking's territory.
- Cadence
- Weekly: a reputation moves slowly, and the email alert covers the interval. Switch to daily for the length of a crisis or a launch, then come back down; the change opens a new version of the tracker and the series carries on.
- Resolution
- hd to set the baseline. Go up a notch when a deviation must be a certainty before you act: resolution repeats every prompt at every check, and that repetition is what turns presence into a stable rate rather than a coin flip.
Reading: a baseline first, then the deviations
The first weeks establish your normal: answer rate per engine, tone distribution, the facts the answers repeat.
After that, three things deserve your attention: a negative wording that appears, an answer rate that drops on one engine, and a wrong fact, an old address, a discontinued product, a departed executive. Open the full answers: every figure traces back there, and the source cited next to the faulty passage tells you where the story comes from. That source is what you are going to fix.
Fix at the source: the canon, then your pages
You do not correct an AI answer inside the engine: you correct the pages it reads. The reference wording is written once, in the canon.
Put in the canon what the answers should say: one-liner, short and long descriptions, category, identity facts. Then align every page where you have the final say: the surfaces registry carries a checklist per page type, from the LinkedIn profile to the directory listing down to your site's llms.txt, and the alignment date freezes into the journal. When the canon is revised, the registry names the outdated pages on its own.
The Atlas and corroborations: a reputation is proven away from home
Engines believe what third-party pages repeat. The map of those pages exists: it is the Atlas, narrowed to your project.
Candidates do the first sorting for you: the pages where an engine already showed your name in its own answers. Record the real ones as corroborations, exact page and publication date; verification rereads the page and archives the passage around your name. Arm monitoring, weekly, on the few pages that carry your reputation, a review page, a reference article: you will know the week the passage changes, the link drops, the page dies.
Quests and Logbook: the routine of a well-kept reputation
The discipline of this use fits in two gestures: date what you do, and let the file tell you what to do.
In the Logbook, record what could explain a curve: a press release, a site overhaul, a public reply to a review, a press campaign. A tone dip then reads with its cause in front of it. In Quests, the file fills itself: pages to realign after a canon revision, watched mentions to reread, candidates to sort; add your own errands, get a press page refreshed, get a listing corrected.
Worked example
A negative answer appears on one engine
The email alert flags it. Open the full answer: the engine cites an old forum thread. Record that thread as a corroboration to keep an eye on it, obtain a fresher page that states the real situation, two recent reviews, an article, log the publication in the Logbook, and let it run: the next checks tell you whether the engine switched sources. That is the whole loop of this use: measure, fix elsewhere, measure again.
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 reputation tracking for [my brand]: four customer prompts about my name, my name as favorite keyword with its variants, Presence and Sentiment analysts, weekly in hd. Then read me the baseline and flag any negative answer.
$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.
Your keywords evolve with you
Add an acronym, a trade name or a domain as a keyword at any time: its series starts at the next check. Your past answers stay archived and readable, nothing is lost.
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.