Engine

Architecture

How the pieces connect

Engine is one pipeline, not a pile of separate tools. Everything it measures feeds one score; everything that score cannot explain becomes a diagnosis; every diagnosis carries the fix that resolves it, applied and verified in place. Click any step for detail.

  1. 01

    Measure

    Search, AI answers, competitors and the web's mentions of you — sampled repeatedly, never asked once.

    • Knowing where you rank in Google — across every market, device and language you care about, and the moment it moves. Rank tracking is not the interesting part of Engine, and it is not meant to be: it is the search half of your visibility score, and one of the main things that sets off a diagnosis. It has to be correct and fast rather than clever.

      See this feature →
    • Answer engines — ChatGPT, Perplexity, Gemini, Copilot, Google AI Overviews — increasingly decide what people discover. You need to know whether you appear in those answers, how accurately you are described, and how your share compares to everyone else. Engine reports this as a confidence band from repeated sampling, never as a single false-precise number, because one query to one model on one day is not a measurement.

      See this feature →
    • Before you can track anything you have to decide what to track — both classic keywords and the prompts people actually type into AI engines. This is the on-ramp: it populates rank tracking and the prompt bank behind AI visibility. Prompts are treated as first-class research objects, and the research covers vernacular, Indic-script and transliterated queries rather than assuming English.

      See this feature →
    • Visibility is relative — the useful question is where competitors beat you, and why. Engine compares search and AI answers together rather than in two separate reports, spanning keywords, citations, content and links at once. "They outrank us" and "they get cited and we do not" become two readings of the same comparison.

      See this feature →
    • For AI visibility, unlinked brand mentions carry more signal than backlinks — but most tooling indexes links and ignores mentions. Engine tracks both, and scores which domains the answer engines actually draw from, so you get a ranked list of places worth earning a mention rather than an undifferentiated backlink export.

      See this feature →
  2. 02

    Score

    One blended number, weighted by your own traffic mix, reported as a range rather than a false-precise figure.

    • Rank tracking, AI visibility and local each answer a fragment of one question: how discoverable are you? Your visibility score blends search, AI and local into a single measure, weighted by your own traffic mix rather than a generic average. One number, then depth — and the number carries its confidence range with it.

      See this feature →
  3. 03

    Diagnose

    When the score moves, or on its own schedule, Engine checks why — technically, on the page, across the web, and locally.

    • Engine reads your site the way a browser does, finds what is technically wrong, and — the part that matters — gives you problems that can actually be fixed rather than a list of warnings. Alongside the standard technical checks it asks whether AI crawlers like GPTBot, ClaudeBot and PerplexityBot are even permitted to read you at all. Most audits never ask, and a surprising number of sites are quietly blocking them.

      See this feature →
    • Whether an AI engine quotes you depends largely on whether your page is easy to lift an answer out of: answers up front, passages that stand on their own, real coverage of the subject, credible signs of expertise. Engine scores that page by page — and then drafts the rewrite, rather than leaving you with a number and no next step.

      See this feature →
    • AI engines and Google's Knowledge Graph reason about businesses, not pages. Engine checks whether yours is described consistently everywhere it appears, whether your own site correctly identifies who you are, and whether your Knowledge Panel is accurate — then fixes the gaps. It is what lets search and AI be one picture instead of two.

      See this feature →
    • Large sites need to know how crawlers — including AI bots — actually spend their time, and which pages are never reached at all. This is a long-standing capability for big sites; the addition here is treating AI bot traffic as a first-class part of it.

      See this feature →
    • For anything with a physical presence, local visibility is not a side report — it is a large share of how you are found. Engine checks your Google Business Profile, whether your name, address and phone agree everywhere they appear, and the state of your reviews, then folds local into your overall visibility score. Findings can be fixed on your Business Profile directly.

      See this feature →
  4. 04

    Fix

    Every finding becomes a change you can approve — applied, checked, and reversed automatically if it does not land.

    1

    Fix Queue

    Every problem becomes a specific, reviewable change — never a report you are left to act on alone.

    2

    Apply

    You approve it once. Engine makes the change directly, in your site, your CMS or your Business Profile.

    3

    Verify

    Engine checks the fix actually landed — not just that it was sent.

    Landed

    Confirmed live, with a record of the change.

    Rolled back

    If it does not land clean, Engine reverses it — no manual cleanup.

  5. 05

    Where it lands

    No new place to manage. Changes land where you already work.

    Every applied fix goes straight to the surface it belongs on:

    Your site Your CMS Your Business Profile