Skip to content

AI visibility, measured

What AI tells your buyers about you

Sondar records how ChatGPT, Perplexity, Gemini and Google AI Overviews answer the questions your buyers ask, verbatim and timestamped, then changes the answer and proves the change stuck.

The offer

Three stages, sold in sequence

An evidence-transparent audit, a fix package that changes the answer, and a retainer that proves it stuck. Monitoring is never sold without the work attached.

  1. 1. See what the engines say, verbatim.

    Audit

    A full Panel Run across ChatGPT, Perplexity, Gemini and Google AI Overviews for your vertical. You get a Visibility Score shown as a band, Share-of-Voice against the competitors you name, the sources the engines cite, and every answer verbatim and timestamped.

    Price$500 – $2,500

    What you get

    • Buyer-intent Prompt Panel for your vertical
    • Per-Engine results with the evidence behind every number
    • Prioritised fix package drawn from the Recommendations Library
  2. 2. Change the answer.

    Fix

    We execute the fix package on your site and Google Business Profile: the pages, structured data, pricing and FAQ content the engines were missing or citing from a competitor.

    PriceScoped from the audit findings

    What you get

    • Content and structured-data changes on your own properties
    • Each change logged against the gap it closes
    • No llms.txt or other placebo work
  3. 3. Prove the change stuck.

    Retainer

    Monthly Panel Runs against your baseline, with deltas reported inside published Variance Bands so a real move is never confused with engine noise. Monitoring is only sold with execution attached.

    Price$1,000 – $2,500 per month

    What you get

    • Monthly report with banded deltas per Engine
    • Continued execution on the open gaps
    • Evidence export on request

Pricing anchors. The exact quote follows the free scan and the strategy call.

Why it is different

A laboratory report, not a dashboard

Verbatim evidence
Every score expands to the exact, timestamped answer behind it. Nothing is summarised away.
Bands, not points
AI engines are non-deterministic, so results are reported inside published Variance Bands. A real move is never confused with noise.
One vertical at a time
Prompt panels are mined for your industry and the fix package respects its claims rules. What works for SaaS does not work for a clinic.

Start with the free Teaser Scan

Ten buyer-intent prompts across all four engines, run once, with two or three verbatim excerpts back in your inbox. No dashboard and no account: an operator runs it and sends you the report.