laup.ai wordmark: "laup" in flowing calligraphic script with connected letterforms, followed by ".ai" in clean sans-serif.
Log inSign up
Methodology

A method you can inspect.

Laup separates what AI said, what the data shows, what models infer and what the site scan finds. These pages publish the rules, evidence boundaries and arithmetic behind each result.

The evidence map

From question to verifiable work.

The paid product follows one evidence chain. Readiness sits alongside it as a technical diagnosis, not a prerequisite or visibility score.

Main evidence chain

AI Visibility and AI Optimization share this sequence. Each result can be traced back towards the captured answer.

  1. 01

    Buying questions

    Controlled by you and tailored to the categories, market and language that matter.

  2. 02

    Engine answers

    Captured from each configured AI engine on a fixed tracking schedule.

  3. 03

    Names and context

    Mentions, citations and retrieved grounding sources are recorded separately.

  4. 04

    Categories and trends

    Visibility is computed by engine, category and time window instead of one hidden score.

  5. 05

    Evidence-backed Actions

    Models propose work; references are verified and displayed numbers are computed in code.

  6. 06

    Measure again

    The same affected area is observed after implementation to see whether the signal changed.

Separate diagnostic

Site readiness

Checks whether the site can be fetched and understood through crawler access, discoverability, rendered content and structured data. It informs the diagnosis, but it does not decide whether AI recommends you.

Three methods

Use the method that matches the question.

Each method has a different evidence base and a different limit. Keeping those limits visible is part of the result.

AI Visibility

What does AI say?

Read the AI Visibility methodology

Measures

Captures real engine answers, who is named, who appears instead, citations, grounding context and change across engines and categories.

Does not establish

Does not establish why an engine chose a company, what a buyer did next or how much revenue AI caused.

AI Optimization

What should change?

Read the Action plan methodology

Measures

Combines measured evidence with independent model analysis. References are checked against Laup data, while counts and competitor gaps are computed in code.

Does not establish

Does not promise that a recommended Action will cause a specific result or move on a fixed schedule.

AI readiness

Can AI access and understand the site?

Read the AI readiness methodology

Measures

Inspects crawler access, discoverability, rendered content and structured product information at one point in time, then applies fixed scoring weights.

Does not establish

Does not measure whether AI names or recommends you. Readability can affect visibility, but it is not the visibility result.

Shared integrity

Rules that apply across Laup.

Different methods produce different outputs, but they share the same discipline around source material, scope and claims.

  1. 01

    Record before interpreting

    Raw engine responses and site observations are captured before scores, trends or recommendations are derived from them.

  2. 02

    Keep the scope attached

    Engine, category, market, time window and the relevant denominator stay with the finding instead of disappearing behind one blended number.

  3. 03

    Label the evidence boundary

    A citation is used or pointed to in the final answer. A grounding source was retrieved into the answer context. Neither alone proves why a recommendation happened, and movement over time is a signal rather than automatic attribution.

Start with the observation

Begin with what AI actually said.

The AI Visibility methodology shows how questions become captured answers, category measurements and evidence your team can inspect.

Read the AI Visibility methodologyStart free trial
Methodology | Laup