Research infrastructure
A reproducible research platform for studying how generative AI systems select, cite and recommend organizations, brands and digital sources.
01
Generative search changes digital visibility from ranked links to synthesized answers. Organizations can be mentioned, cited, recommended or omitted, but current practice still lacks sufficiently standardized and reproducible ways to measure those outcomes.
02
From SEO to GEO (Generative Engine Optimization): development and validation of a scientific model to optimize organizational visibility in AI-based generative search engines.
Research workflow
GSLHub is designed so that every scientific result can be traced backwards to its experimental conditions and preserved raw evidence.
Define the visibility question and the scientific gap.
Specify a testable relationship or expected effect.
Freeze prompts, AI systems, targets and repetition rules.
Run controlled prompts and preserve the execution context.
Store response exports, screenshots and provenance with integrity checks.
Calculate governed indicators such as AIR, CR, MCP and RCR.
Retain versioning, review, audit and immutable scientific snapshots.
Methodology
Experiments can reserve multiple comparable executions under the same versioned protocol.
Prompt, AI system and environment context become protected once an execution enters the governed workflow.
Raw exports and screenshots are linked to the execution and can be verified by SHA-256 against persistent storage.
Coding, evidence and metric definitions support quality-control and reviewer separation before validation.
Measurement model
The first GSLHub metric set formalizes whether a target appears, is cited, where it is cited and how consistently the result is reproduced.
Share of eligible responses in which the target is mentioned.
Share of eligible responses in which the target is explicitly cited.
Average observed position of target citations in eligible responses.
Rate at which comparable repetitions reproduce the coded outcome.
Development pilot
GSL-EXEC-GEO-0001 is a development validation run used to prove the research chain before any doctoral data collection begins.
This pilot is development evidence, not a doctoral result. A Final Development Reset will separate all development records from future doctoral research data.
Execution
Completed / Published
Raw response artifact
SHA-256 verified
Screenshot artifact
SHA-256 verified
Evidence records
2 validated
Observation
Validated / Published
Visible citations
0 observed
Reproducibility
The goal is not only to calculate a metric, but to make it possible to explain where that metric came from and whether the underlying evidence remained unchanged.
The public dashboard exposes only publishable research indicators. Authorized researchers use the private CMS for governed research operations.