Frequently asked questions

Whether AI search engines mention or recommend a brand for the prompts you track, which competitors appear instead, and which sources the engines cite. Results are summarised as the Norund Score, share of voice, and citation intelligence.

The demo core tracks OpenAI (search-grounded API) and Perplexity (Sonar API), both via official APIs. More engines, starting with Google Gemini, are planned via the same adapter architecture.

Not always. We use official APIs, and consumer interfaces add their own context, memory, and personalisation. AI answers also vary naturally between runs. Treat results as a representative visibility signal, not a screenshot of one user's chat.

A deterministic 0–100 score computed by documented code (never by an AI model) from how each engine answered each tracked prompt: position and type of mention, adjusted by sentiment, averaged per engine, then blended with equal engine weights.

Across all valid responses in a scan, we count appearances for the client brand and each tracked competitor and divide by total valid responses. Direct recommendations and incidental mentions are reported separately.

The domains and pages engines cite when answering your tracked prompts. Frequently cited sources are evidence signals — citation does not guarantee recommendation.

Weekly on all plans, plus manual scans with a per-plan cooldown. Daily scanning is a planned add-on entitlement.

Yes. Every completed scan can be exported as a PDF report with the agency's logo, accent colour, and introduction.

On white-label plans the PDF carries your agency branding and no platform branding, so you can deliver it directly to clients as part of your own service.

No. Nobody can. The platform measures visibility signals and cannot guarantee rankings, citations, traffic, or revenue.

AI engines are probabilistic and their indexes update constantly. Tracking that movement over time is precisely what the platform is for.

Yes — prompts are free text, so 'best ecommerce SEO agency in London' works. Market and language fields help you organise them.

Up to three active competitors per workspace in the current build.

Row-level security isolates every organisation's data; provider API keys stay server-side; public-scan IPs are stored only as salted hashes; and raw engine responses are only visible to your organisation.

Workspaces can be archived and restored at any time. Full account deletion is part of the production-launch checklist before billing opens.

The scan finishes as partially completed. Failed calls are excluded from scoring — they are never counted as a zero — and the dashboard shows exactly what failed.

Yes — five prompts across two engines for any domain, with an email address. It produces a partial score and one competitor or citation insight.

Yes, results are cached for 24 hours per domain. Refreshing the page or resubmitting reuses the cached result instead of triggering new engine calls.

No. The current product measures visibility; content generation is not part of it.

Gemini tracking, scheduled client email reports, daily scans, additional engines, BYOK support, and more — see the pricing page for planned add-ons. We only list features here once they exist.