Measure and improve how AI answers represent your brand

Welcome

Searchestra measures and helps improve how brands appear and compete in AI answers. Instead of guessing how ChatGPT or Gemini talk about your brand, you get a measured, repeatable picture across AI platforms, the competitors recommended instead of you and the actions that close the gap.

• Organized like your business: organizations, brands, markets and Prompt Universes.
• A transparent method: a versioned, brand-neutral question set, so trends are comparable.
• Honest data: we report only what we actually collect.

How Searchestra is organized

Everything in Searchestra sits in one hierarchy, so a result always has its context:

• Organization: your company account. It holds your brands and your team.
• Brand: a brand you measure. A brand can be measured in several markets.
• Market: a country and language, and optionally a product line or category, where the brand competes. Competitors are tracked per market.
• Prompt Universe: one buyer need or decision context within a market, measured with its own versioned question set.

The same brand can lead in one Prompt Universe and be absent in the next, which is why results are always read in this chain: brand, market, Prompt Universe, questions, answers.

Create your account

1. Go to panel.searchestra.com/signup and sign up with your email.
2. Searchestra creates your organization. If you signed up with a company email, we pre-fill your domain; with a personal email (like Gmail), you enter your real domain during onboarding.
3. You land in the app, ready to add your first brand.

Onboarding

Onboarding sets up everything the measurement needs, in a short wizard.

1. Enter your domain. Searchestra reads your site to learn your brand, category and typical topics.
2. Set up the brand and market. Review the detected brand name, aliases, target country and language, and the competitors we found; edit anything.
3. Build the Prompt Universe. We generate a brand-neutral, IAB-aligned set of buyer questions for the market.
4. Invite your team (optional), then run the first scan.

Run a scan

Searchestra re-measures on a regular cadence. For an on-demand update, open a Prompt Universe, go to Overview and click Scan now. A scan collects fresh answers for the question set across the enabled AI platforms, then recomputes every metric. A progress bar shows how far along it is.

Large question sets across several platforms can take several minutes. You can leave the page; the scan keeps running and results appear when it finishes.

AI platforms

Our catalog has nine AI platforms, collected with the same method so results are comparable run to run:

• ChatGPT · Claude · Gemini · Perplexity · Grok · DeepSeek · Meta AI
• Google AI Overviews and Google AI Mode

Web search. Each platform is asked with its provider’s own web search switched on wherever the provider offers it (ChatGPT, Claude, Gemini, Grok and Meta AI). The model decides per question whether to search; Searchestra never forces a search. An answer given without a search is a real observation of how that platform behaves and stays in every metric. DeepSeek’s API offers no web search, so its answers come from the model’s own knowledge. On Perplexity, AI Overviews and AI Mode the search is part of the product rather than the model’s choice: Perplexity’s answer setting instructs a search before every answer, and AI Overviews and AI Mode are Google Search features that build the answer from several related searches. Google shows an AI Overview only when it judges it helpful, so some questions have none. Searchestra records those questions and shows how often an AI answer was shown (the AI answer shown column on the Citations overview); they do not count as answers in the other metrics.

Active coverage for a given run depends on which platforms are enabled and available for your account, and each run shows exactly which were included. If a platform or data provider is temporarily unavailable, Searchestra records that and narrows the affected metric for that run. It never invents a response to fill the gap.

Prompt Universe

A Prompt Universe is one buyer need within a market, such as small kitchen storage in Türkiye, together with the questions Searchestra asks AI platforms to measure it. The question set is brand-neutral (it doesn't lean on your own marketing), versioned and reproducible, so your visibility is measured against a stable, defensible baseline rather than a single lucky prompt.

Each question carries its intent (informational, commercial, comparison, transactional or navigational) and scenario. Versions keep historical comparisons honest.

Deep dive: why the prompt set is the single biggest driver of your numbers.

Universe overview

Each Prompt Universe opens on its Overview. At a glance you see:

• Answers collected: how many AI answers this period is based on.
• AI Visibility: the share of answers that mention you.
• Competitor analysis: your position and share against the leaders.
• Citations: how often your pages are cited.
• IAB 4P panel: Presence, Prominence, Portrayal and Persuasion.

Brand and market overviews roll these up so you can compare markets side by side, without averaging different markets into one number.

Deep dive: what GEO is and how AI platforms decide which brands to name.

Metrics & the IAB 4P model

Searchestra reports on the IAB 4P framework, plus supporting KPIs.

Deep dive on each metric: the 4 P's, mention rate, share of voice and citation rate. See also how we measure, our full methodology in the open.

Competitors

Competitors are tracked per market. Searchestra detects the brands AI platforms actually name in your category and ranks them into a leaderboard, so you see who leads, your position and the gap in visibility, share of voice and citations. Add or remove competitors any time; the next scan reflects the change.

Citations & sources

When AI platforms cite sources, Searchestra records which domains and pages appear as citations in your category. Use this to find where a link to you is missing, which third-party pages are cited alongside competitors, and where earning a citation would close a gap.

A citation is a source the platform itself returns with its answer, so it appears only when the model searched. The Web search by platform table on the Citations overview shows who decides the search on each platform and, where the model decides, how often it chose to search, and an answer’s detail page shows whether that answer searched and how many times. An answer without citations still counts: it did not cite you, which is a real outcome rather than missing data.

Deep dive: citations vs mentions and source authority in AI answers.

Action Center & playbooks

The Action Center turns findings into a ranked list of work: the content, claims and sources behind your largest observed gaps, prioritized by the evidence we collected. It exists at universe, market and brand level. Playbooks group related actions into a repeatable plan, so a team can work a gap systematically and re-measure on the same questions.

Our honesty principle

Searchestra never fabricates measurement data. If a provider fails, we say so and narrow the metric for that run. A number you see is a number we collected, which is what makes trends trustworthy enough to act on.

We publish the full method in the open: how we measure. Deep dive: directional vs decision-grade measurement, and why AI visibility tools disagree.

Team & access

Access is set per brand. Invite teammates to the brands they work on and give each a role: Brand admin, Editor or Viewer. Members only see the brands they have access to. For how we protect your data, see Security & Trust.

Data portability

You can take your organization's data with you at any time. An owner starts an export in Settings and downloads one package: records as JSON (large tables also as CSV), files such as reports in their original format and a manifest with a checksum for every file.

To move to another provider or to your own systems, start a switching request in the same place. You choose the notice period, the switch takes at most 30 days after it, and your data stays available for download for at least 30 more days. We do not charge for switching.

The full list of what the export contains, with every file and field, is published in the data portability register. It is also available as JSON for your own tools.

FAQ

Why is my visibility 0%?

Because, this period, AI platforms didn't mention your brand in answers to the category questions. It's an honest baseline; the Action Center shows how to earn mentions and citations.

How often is my brand measured?

AI answers drift, so Searchestra re-measures on a regular cadence to make trends meaningful. You can always click Scan now for an on-demand update.

Can I track more than one brand or market?

Yes. An organization can hold several brands, each brand several markets and each market several Prompt Universes, each with its own scans, competitors and history.

Does clicking "Scan now" many times run many scans?

No. A Prompt Universe runs one scan at a time; extra clicks are told it's already in progress, so you never get duplicate runs.