HomePrivacy

Measurement is our product. Your data isn’t.

Praised is a measurement service that reads the public web, not you. This policy covers what we collect, who else processes it, and how long we keep it — with no tracking cookies and no third-party analytics script anywhere on this site.

Who we are

Praised is an AI-visibility measurement service: it asks AI tools like ChatGPT questions about your category and reports how they answer. It is operated by Praised Oy (Business ID 3454899-8, Kivenhakkaajankuja 3 B, 02650 Espoo, Finland), which is the data controller for the information described below. Questions, access requests and deletion requests go to hello@praised.co.

Visiting this website

We count page views to know which pages are worth keeping. That counting is done by a small script of our own, which sends the page address and the referring site to our own server — no cookie, and no third-party tag.

For each page view we record the path, the referring site's domain, whether the device is mobile or desktop, and a visitor hash: a salted hash of your IP address and user-agent string that is re-salted every day. The raw IP address is never written to our database, and because the salt rotates daily, the same visitor hashes differently tomorrow. It can count how many people read a page; it cannot follow a person from one day to the next. Raw page-view rows are deleted after 90 days; only aggregate counts are kept longer.

We store two things in your browser's local storage, and neither is used for tracking: your sign-in token, and your light/dark theme choice. Inside the signed-in app only, PostHog (listed below) also keeps an anonymous id and session state there, with no cookie; it is never loaded on this website.

Using the free tools

The free checker and AI-access check take a public URL, fetch it the way an answer engine would, and score what they find. The result is stored so it can be shared as a link, and it expires. If you ask us to email you the report or watch a site for changes, we keep the email address you gave for that purpose and nothing else, and every message we send carries a one-click unsubscribe we honour permanently.

We also keep a record of what was asked: the search or URL you typed, the domain you were checking, which tool it was and whether it ran. We use it to see which questions people bring us — so we know what to write about and what to build — and, where the domain is a company we could help, to decide who to contact. This is a company-level record: we do not store your IP address, a visitor identifier or your browser alongside it, so it says a business looked something up, not who you are.

The lawful basis is our legitimate interest in understanding demand for a business product, balanced against the fact that no individual is identified. We keep these records for 18 months and then delete them. You can ask us to remove entries for your domain, or object to the whole thing, at hello@praised.co — and we do not share any of it with advertising networks or use it to build ad audiences.

Having an account

  • Identity — your email address, name if you give one, and your workspace. Passwords are stored only as a PBKDF2-SHA256 hash with a per-user salt; we never see or store the password itself. If you sign in with Google or your company's single sign-on, we receive your email address from that provider and no password at all.
  • Billing — plan, credit balance and invoice history. Card details are entered directly with Stripe and never reach our servers.
  • Product usage — which features an account used and when, so we can see what works. Stored against your workspace and, where it matters, your email as the actor.
  • Email summaries — New Praised workspaces get a weekly email summary of what changed in how AI assistants describe them, sent only in a week when something changed. Anyone who receives it can turn it off with one click in the email, or in Settings → Notifications. Workspaces created before this change keep the choice they had.
  • Your measurement data — the brands, questions and pages you track, the readings computed from the answers, and the findings and content you generate. Your brand, your question list, your readings and every number on them are scoped to your workspace, and no other customer can see them. Answers are stored differently, and it is the one thing on this page worth reading twice — see Where answers come from below. Two features do compute across workspaces, and only ever over de-identified values: the peer benchmark, which tells you your percentile against other measured projects, and source authority, which counts how often a domain is cited across the projects we measure. Both release aggregates only — medians, quartiles, counts — never a row, never a brand, and never anything attributable to a workspace, and both withhold a result entirely below a minimum cohort size.

Where answers come from

A question like “best contract analysis software” is a public question, and the answer an AI tool gives to it is not about you. We store each answer once, keyed by the exact request that produced it — the question text, the engine, the model, the place and the language — and any workspace tracking that same question is served that same answer.

Nothing about you travels with an answer. An answer carries no workspace, no organisation, no brand and no record that you were the one asking. There is no customer identity attached to it to remove, because none is ever attached. The console shows you the answers your questions were served and never shows where an answer came from or who else is tracking it — not because we hide it, but because that information does not exist in the store.

What this buys is not history. A question several workspaces track is asked once rather than once for each, and nothing on the answer says who it was asked for. Your chart still starts on the day you approve a question — not because we reach into the library's past, but because a newly tracked question is asked straight away, and that first answer serves only you.

  • An answer counts for you only from the day you started tracking that question. We do not reach backwards into the library's history on your behalf.
  • Questions you mark private are never shared, in either direction. On Enterprise you can mark a question private: it is never answered from the shared store, and the answers it produces never serve another workspace. The isolation runs both ways by design, not by policy.
  • Answers asked with your own provider key belong to your organisation alone and never serve anyone else.
  • Answers from the free tools on this site are kept separately and are never used to answer a customer's tracked question. The free checker always asks fresh.
  • Your brand profile is read from your own answers. Which brands an answer mentions is decided per workspace, against your own profile, and that reading is never shared.

What we send to AI providers

Measuring visibility means asking real AI tools real questions. The prompts we send are about brands, products and categories — the questions your buyers ask. We also send the text of pages fetched from the public web — yours, or a URL you enter — when you run a page audit, build a brand profile or draft content, because those features work by reading the page. We do not send your customer lists or your account details, and we do not use provider APIs to process personal data on your behalf.

Who else processes your data

SubprocessorWhat for
RailwayHosting, application databases and encrypted backups
StripePayments, invoicing and tax. Card data is held by Stripe, not us
ResendTransactional email — sign-in, verification, invitations, reports and alerts
Anthropic, OpenAI, Google, Perplexity and other answer enginesThe measurement itself: the prompts we ask on your behalf
SearchAPI.ioReading Google AI Overviews and AI Mode — the search terms we run on your behalf, and the ones typed into the free AI Overview checker
SerpApiThe same reading, when SearchAPI.io cannot answer
GoogleSign-in with Google, and web fonts on our marketing pages
PostHog (EU Cloud, data stored in Frankfurt)Product analytics and session replay for signed-in app users. Receives a pseudonymous account id, the workspace's id, URL name and plan, the pages viewed and the clicks made, and session recordings that show a page's layout and clicks only: all text, every form input and every label, title and link address are masked, so no AI answer text, email address, name, workspace name, key or secret is recorded. Not used on the public site, the sign-in page or shared reports, and not loaded when the browser sends Do Not Track or Global Privacy Control

We will update this list before adding anyone to it.

How long we keep things

  • Free-tool results: they expire, and a shared link stops working when they do.
  • What was asked in a free tool (the query, the domain, the tool): 18 months, then deleted. No IP or visitor identifier is kept with it.
  • Raw page-view rows: 90 days, then deleted. Aggregates have no personal data in them.
  • We keep a record of which product emails and in-app notices we sent, delayed or skipped for each person, without their contents, for 180 days.
  • Account and measurement data: for as long as the account exists. Ask us to delete it and we will, including from the backups on their next rotation.
  • Answers: 400 days, compressed, then deleted. A reading is kept for as long as the account exists, with a version for each time it was computed — adding a spelling to your brand profile re-reads the days behind you, and the earlier version is kept rather than overwritten.
  • Free-tool answers: 30 days, then deleted. They are never used to answer a tracked question.
  • Backups: encrypted, retained on a rolling schedule for disaster recovery only.

What deleting your workspace removes

Everything that is yours: your questions and their wording history, your readings and every figure on them, your brand profile, your competitors, your pages, your drafts, your Actions queue, your account and your billing history. Also the links that record which answers your questions were served — after which nothing connects you to any answer at all.

An answer to a public question that another workspace is also being served is not deleted, because it was never yours and carries nothing of yours. An answer no other workspace is served is deleted with the rest. If you would rather nothing you asked outlives your account under any circumstances, mark your questions private before you start: a private question's answers serve you alone, and go when you do.

Your rights

You can ask for a copy of your data, ask us to correct it, ask us to delete it, restrict how we use it, take it elsewhere, or object to it altogether. Email hello@praised.co and we will answer within 30 days, free of charge.

Our lead supervisory authority is the Finnish Data Protection Ombudsman (Tietosuojavaltuutetun toimisto, tietosuoja.fi). If you think we have got something wrong you can complain to them, or to the authority in your own EU or UK country — you do not have to come to us first.

Security

Data is encrypted in transit. Access is scoped per workspace and enforced on every request; API credentials are held only as hashes and can be revoked by you at any time. We will tell affected customers about a breach that puts their data at risk without waiting to be asked.

Changes

If we change this policy in a way that matters, we will email account holders rather than quietly updating the date at the top.

See also our terms of service, and the measurement methodology, which explains what our numbers can and cannot tell you.

Questions, answered.

Something else on your mind? Contact us.

You can ask for a copy of your data, ask us to correct it, delete it, restrict how we use it, take it elsewhere, or object altogether. Email hello@praised.co and we answer within 30 days, free of charge. Our lead supervisory authority is the Finnish Data Protection Ombudsman. See Your rights above for the binding wording.

Free-tool results expire, and a shared link stops working when they do. What was asked — the query, the domain and the tool — is kept for 18 months and then deleted, with no IP or visitor identifier stored alongside it. See How long we keep things for every retention period.

They can be served the same answer to the same public question — an answer is stored once and reused — and nothing about you goes with it: not your workspace, not your brand, not that you asked. What stays yours is your brand, your question list, your readings and every number on them. On Enterprise you can mark a question private, and then neither direction happens. See Where answers come from above for the binding wording.

Everything that is yours goes, including the links recording which answers your questions were served. An answer another workspace is also served stays, carrying nothing of yours; an answer nobody else is served is deleted with the rest. See What deleting your workspace removes.

The questions themselves, which are about brands, products and categories — the questions your buyers ask. We also send the text of pages fetched from the public web: yours, or a URL you enter. A question two workspaces both track is sent once, not twice. See What we send to AI providers for the full description.