Per-property onboarding

Onboarding is what turns a bare URL into a tracked property with a first score. It is the same pipeline whether it runs for the first property in a workspace or a new one added from the property switcher — enter a URL, and Lokrix does the rest.

The onboarding funnel

Entering a URL kicks off an ordered sequence. Each step feeds the next, and the property is created before any analysis begins so all results attach to it.

The property onboarding funnelA narrowing five-stage funnel: add property, readiness pre-check, analyze brand, first scan, and report, connected top to bottom.Add propertyReadiness pre-checkAnalyze brandFirst scanReport

1. Enter a URL → create the property

Lokrix normalises the URL and creates a new property inside the current workspace. From this point everything — prompts, readiness, scan and score — is scoped to that property.

2. Analyse the brand

Lokrix crawls the site to detect the brand, industry and the topics buyers associate with it. This context is what makes the generated prompts realistic rather than generic.

3. Generate the prompt universe

From the brand analysis, Lokrix generates the property’s prompt universe — the real questions your audience asks AI engines. These prompts define what every future scan measures for this property.

4. Readiness pre-check

Before spending credits, Lokrix runs a readiness pre-check to confirm the property, prompts and engine configuration are in shape to scan — so the first run does not waste credits on an avoidable failure.

5. Run the first scan

Finally, Lokrix scans the property. Five engines run live by default — ChatGPT, Claude, Perplexity, Grok and Gemini — while Google AI Overviews and AI Mode need a search-provider key to observe. Each prompt is run K times across temperature, persona and locale to produce a Monte-Carlo presence probability with a 95% confidence interval — never a single-run yes/no. The result is the property’s first Lokrix AI Score.

A note on cost

Live runs meter against the workspace credit balance — roughly one credit per engine response, multiplied across prompts, engines, samples and cadence. Lokrix defaults to the predictive model and escalates to live runs deliberately; see Live vs predictive and Credits & cost.