01
AI Discovery
Observe when, where, and under which query conditions businesses surface in generative AI responses.
DISCOVERYVISIBILITYQUERY INTENT
AMIP // APPLIED AI RESEARCHRESEARCH IN PROGRESS
MAGNETIC
SEARCH AI.
AMIP, the AI Market Intelligence Program, is a proprietary applied research initiative studying how generative AI systems discover, interpret, evaluate, and recommend businesses.
AMIPAI Market Intelligence Program
Magnetic Search AIresearch framework
Ongoingapplied empirical research
INTELLIGENCE LAYERAIDISCOVERY + RETRIEVAL
WEBEvidence
IDEntity
LLMRetrieval
RECRanking
AI VISIBILITY◆ENTITY INTELLIGENCE◆RETRIEVAL INTELLIGENCE◆CROSS-MODEL ANALYSIS◆EVIDENCE SIGNALS◆AI MARKET INTELLIGENCE◆AI VISIBILITY◆ENTITY INTELLIGENCE◆RETRIEVAL INTELLIGENCE◆CROSS-MODEL ANALYSIS◆
01 / RESEARCH QUESTION
What makes a business visible to generative AI?
AMIP examines the relationship between digital identity, evidence sources, market context, retrieval behavior, and the recommendations produced by generative AI systems. The objective is to distinguish repeatable patterns from assumptions.
02
Entity Intelligence
Study how business names, domains, locations, categories, and external references are resolved into a coherent commercial entity.
ENTITYIDENTITYRESOLUTION
03
Evidence Signals
Analyze the sources and consistency signals that appear alongside successful discovery and recommendation outcomes.
SOURCESEVIDENCETRUST
04
Cross-Model Behavior
Compare outcomes across generative AI systems to identify differences in retrieval, source affinity, context, and recommendation behavior.
LLMRETRIEVALCOMPARISON
02 / MAGNETIC SEARCH AI
A research framework built around observation and re-testing.
Magnetic Search AI organizes the AMIP research cycle so that observations can be documented, compared, challenged, implemented, and tested again rather than treated as static SEO assumptions.
Observe → Measure → Compare → Diagnose → Implement → Re-test
The framework records business identity, query conditions, discoverability, source evidence, model behavior, and changes over time. Its purpose is to create a defensible research trail for future analysis and publication.
Publication track ↗Separate observations from hypotheses.
AMIP distinguishes measured outcomes, inferred relationships, and untested hypotheses so research claims can evolve as the dataset grows.
Analyze businesses inside real competitive environments.
The program is designed around real companies, categories, locations, digital evidence, and competitive discovery conditions rather than isolated laboratory prompts.
RESEARCH STATUS
Ongoing applied research. Findings will be published only when supported by the observed dataset and documented methodology.
03 / PUBLICATION TRACK
Built to become a publishable research study.
The AMIP research structure is being organized so that future publications can clearly document the research question, dataset, methodology, observations, limitations, findings, and references.
Research Question
Define the specific discovery or recommendation behavior being tested.
Methodology & Dataset
Document companies, queries, comparison criteria, evidence, time windows, and model conditions.
Findings & Limitations
Report repeatable observations while stating uncertainty and conditions that may affect reproducibility.
Publication
Convert validated work into a structured paper under the AMIP / Magnetic Search AI research initiative.
04 / RESEARCH LEAD
Research connected to enterprise security experience.
AMIP is led by Mario Martinelli through Almeida Martinelli Consulting and Service LLC, extending a 16+ year enterprise technology background in SAP Security, GRC, Identity & Access Governance, risk, and compliance into applied AI research.
16+Years enterprise SAP experience
AMIPApplied AI research program
AIDiscovery + retrieval research
R&DPublication-oriented methodology
05 / CONTACT
Research, enterprise security, and strategic AI conversations.
Contact Almeida Martinelli regarding SAP Security, enterprise governance, AMIP research, or strategic collaboration.
ALMEIDA MARTINELLI