Getting started
Core concepts
Eight terms cover everything in GEOX. Learn these and every page in the dashboard will read naturally.
Brand
The product or company you monitor. A brand carries a name, aliases (comma-separated variants so mentions are caught even when phrased differently), an optional URL and description, and a status — Active or Paused. Paused brands keep their history but are skipped by future runs. You can monitor several brands; the top-bar selector picks which one you’re viewing, and every page scopes to it.
Competitors
The rivals you measure against, each with their own name and aliases. Competitor names are fed into the analysis pass so GEOX can compute your mention share head-to-head and find the questions competitors win that you lose. Manage them on the Brands page.
Questions
The tracked prompts your customers might ask an AI — e.g. “best project management tools?”. These are what the engines actually run. Each question can be toggled Active or Paused; paused questions are skipped on collection runs but keep their history. Questions come from two places: ones you write by hand on the Brands page, and mined questions you chose to track in Question Mining.
Topics
An optional tag on each question (e.g. “pricing”, “alternatives”). Topics group questions so analytics can break performance down by theme. Untagged questions roll up under Untagged.
Collection (a “run”)
A run sends your active questions to the enabled engines and stores the answers. Trigger one manually with Run now, run a single question with its ▶ button, or let a schedule fire it automatically. See Collections & scheduling for the full pipeline.
Responses
The raw AI answers, stored verbatim and never deleted — they are the source of truth everything else is derived from. Each response is tagged Manual or Scheduled so you always know where it came from. Browse them on the Questions page or per-question in the answer history.
Metrics
What the extraction pass produces from each answer: whether the brand was mentioned, its rank when the answer is a list, the sentiment of the mention, and any cited sources. Only analyzed answers count toward analytics — empty or failed model replies are stored but excluded. The Metrics reference page defines each number precisely.
Schedule
An optional per-brand cadence — daily, weekly, monthly, or a one-time run — set from the clock icon on a brand row. Scheduled runs execute in your local time zone.
Cheat sheet
| Term | In one line |
|---|---|
| Brand | What you monitor — name + aliases, Active or Paused. |
| Competitor | Who you're measured against in share-of-voice and gap analysis. |
| Question | A prompt the AI engines are actually asked, Active or Paused. |
| Topic | Optional tag that groups questions for thematic breakdowns. |
| Run | One collection: active questions × enabled engines → stored answers. |
| Response | A raw answer, saved verbatim, tagged Manual or Scheduled. |
| Metrics | Mention, rank, sentiment, citations — extracted from each answer. |
| Schedule | Per-brand automatic cadence for runs, in your local time zone. |
Paused vs. deleted
Pausing keeps history and lets you resume later; deleting removes the configuration. Raw responses are never thrown away either way.