Turning complex data into intelligent business decisions.
AI, Business Intelligence and data architecture consulting for companies in Costa Rica and Latin America. Specialists in public procurement intelligence, credit scoring and consumer analytics.
Nancy Rodríguez is an economist, statistician and Data & AI leader with experience transforming complex business problems into measurable, data-driven solutions.
Her career spans economic research, statistical methodology, financial and commercial analytics, consumer intelligence, cloud data ecosystems and artificial intelligence.
She combines rigorous quantitative methods with hands-on implementation — from measurement frameworks and data architecture to predictive analytics, automation and AI-powered decision systems.
Her work combines
Artificial Intelligence
Advanced Analytics
Data Strategy
Consumer Intelligence
Predictive Modeling
Data Governance & Maturity
Executive Intelligence Systems
From public procurement intelligence to consumer behavior analytics, N-AI's focus is not just building dashboards — it is building systems that turn data into actionable intelligence.
Why N-AI
Most organizations already have data.
The challenge is knowing how to structure it, connect it, interpret it and turn it into intelligent systems that support real decisions.
N-AI was created to close that gap.
From research to artificial intelligence
Artificial intelligence is not the starting point of N-AI's methodology. It is the latest layer of a career built on economics, statistics, measurement and business analytics.
1Economic research→
2Statistical methodology→
3Business analytics→
4Data science→
5Consumer intelligence→
6Data & AI leadership
“Data without methodology is just information.”
My background in economics and statistics shapes how N-AI approaches artificial intelligence: starting with the business question, defining the measurement framework, validating the data — and only then selecting the technology.
Complete SICOP intelligence architecture for anomaly detection, risk scoring and preventive monitoring.
Architecture Preview
Sources flow into the pipeline; the pipeline emits intelligence; intelligence powers the executive surface — every component traceable end-to-end.
SourcesSolicitations · Awards · Contracts
Pipeline7 stages · Replayable · Audited
AI Layer47 patterns · Explainable scoring
OutputsRisk index · Triage · Drilldown
Procurement records analyzed
Anomaly patterns modeled
Institutional risk dimensions
Risk scoring latency
01The Challenge
Public procurement operates on opaque ecosystems: fragmented data, inconsistent taxonomies, and oversight that arrives after harm is done.
Auditors, regulators and decision-makers face a compounding problem — millions of transactions across thousands of institutions, encoded in heterogeneous schemas. The result is reactive oversight: anomalies are caught only after irregularities materialize into legal cases, and patterns of systemic risk go unmodeled. The gap is not data — it is intelligence.
Opaque procurement ecosystems
Fragmented, heterogeneous data sources
Inconsistent taxonomies across institutions
Reactive oversight — irregularities surface after harm
Systemic risk patterns left unmodeled
02Data Infrastructure
Inventory and structuring of the complete SICOP ecosystem — every source, every schema, every relationship.
We mapped the SICOP data graph end-to-end: solicitations, awards, contracts, addenda, supplier registries, sanctions, institutional metadata. Each source was inventoried, schemas reconciled, and relationships modeled into a unified graph that downstream pipelines can reason over.
Sources
Solicitations
Awards
Contracts
Addenda
Supplier Registry
Sanctions
Institutional Metadata
Ingestion
• Continuous extraction
• Schema reconciliation
• Lineage capture
• Validation rules
Unified Graph
Canonical schema. Cross-source relationships modeled. Replayable transforms. Substrate for every downstream signal.
03Intelligence Pipeline
Seven-stage pipeline from raw extraction to preventive monitoring — the spine of the whole system.
Each stage is independently testable, observable and replayable. Raw sources flow through extraction, structural inventory, unification, validation, anomaly detection, risk scoring and into the monitoring layer that powers executive surveillance.
Forty-seven anomaly patterns modeled across temporal, structural and behavioral signals.
Rule-based detectors catch the obvious; learned models surface the subtle: unusual award velocity, supplier concentration drift, pricing breaks vs. peer baselines, and addendum sequences that statistically precede irregularities. Outputs are scored, ranked, and explainable.
Signal · Last 60 weeksFlagged anomalies
05Risk Scoring System
Eight institutional risk dimensions composed into a single executive score with traceable components.
Each institution carries a score that decomposes into dimensions auditors and decision-makers actually reason about: contracting velocity, supplier concentration, pricing dispersion, addendum exposure, sanction proximity, transparency posture, control maturity and historical signal density.
8-Dimension Composite
06Institutional Intelligence
A single executive surface where every institution has its own risk profile, peer comparators and trend.
Heatmap-style triage shows the population at a glance; drilldowns reveal time-series, anomaly contributions and historical events. The same surface answers two questions at once: where to look first, and why.
low
INSTITUCIÓN A
23risk
low
INSTITUCIÓN B
31risk
low
INSTITUCIÓN C
28risk
medium
INSTITUCIÓN D
47risk
medium
INSTITUCIÓN E
52risk
medium
INSTITUCIÓN F
58risk
medium
INSTITUCIÓN G
49risk
high
INSTITUCIÓN H
71risk
high
INSTITUCIÓN I
76risk
high
INSTITUCIÓN J
68risk
critical
INSTITUCIÓN K
84risk
critical
INSTITUCIÓN L
91risk
07Strategic Impact
Three executive outcomes the system unlocks — measured in months, not quarters.
Preventive oversight replaces reactive audit. Decision velocity improves because risk is visible at the surface, not buried in records. And transparency becomes a deliverable, not an aspiration.
Preventive Oversight
Risk surfaces before irregularities calcify into legal cases — auditors act on signals, not findings.
Decision Velocity
Executives see risk where it lives — at the surface — and reach a defensible decision in minutes, not weeks.
Operational Transparency
Every score traces back to its components; every component traces back to a record. Transparency becomes a deliverable.
Pipeline
Seven stages from raw to executive surface.
Each stage is independently testable, observable and replayable. The result is intelligence with lineage — every conclusion traces back to the record that produced it.
01
Data Extraction
Pull raw records from every SICOP source on a continuous cadence.
02
Inventory & Structuring
Catalog every field, document every relationship, version every schema.
03
Data Unification
Reconcile heterogeneous schemas into a single canonical graph.
04
Validation Rules
Enforce typing, lineage and referential consistency before downstream use.
05
AI Anomaly Detection
Score 47 anomaly patterns across temporal, structural and behavioral signals.
06
Risk Scoring
Compose 8 institutional risk dimensions into a single auditable score.
07
Preventive Monitoring
Surface flagged events, trends and peer drift to executive decision-makers.
Pull raw records from every SICOP source on a continuous cadence.
Catalog every field, document every relationship, version every schema.
Reconcile heterogeneous schemas into a single canonical graph.
Enforce typing, lineage and referential consistency before downstream use.
Score 47 anomaly patterns across temporal, structural and behavioral signals.
Compose 8 institutional risk dimensions into a single auditable score.
Surface flagged events, trends and peer drift to executive decision-makers.
Demos
Working systems and the ones we're building next.
Production deployments and design-stage prototypes. Each demo answers a real question for a real audience.
N-AI works two fronts with different buyers: institutional and financial intelligence, and commercial and consumer intelligence. The problem changes and so does the counterpart; the quantitative standard does not.
Vertical 01
Institutional & financial intelligence
For audit, risk, finance and the public sector
Anomaly detection & risk scoring
Temporal, structural and behavioral signals modeled at scale, with explainable scoring.
Public procurement intelligence
Analytics on public contracting: patterns, supplier concentration and preventive monitoring.
Credit decision engines
Automated pre-approval, business rules and regulatory compliance with sub-second decisions.
Data governance & architecture
From inventory to lineage, quality, ownership and regulatory posture — strategy made operable.
Vertical 02
Commercial & consumer intelligence
For marketing, commercial and category teams
Consumer intelligence & RFM
Behavioral segmentation, revealed preference and actionable profiles for category leaders.
Media & ad spend analytics
A quantitative read on advertising investment: incrementality, saturation and return by channel.
Social listening with methodological design
Digital listening treated as measurement, not counting: sampling design, validity, and signal over noise.
Data ingestion & unification
Building proprietary sources and unifying data scattered across CRM, campaigns, e-commerce and research.
Areas of work
Real cases, AI prototypes and intelligence systems — the work, in the open.
The areas N-AI works in and publishes methodology on. No promised dates: each piece ships when it is ready.
Case Study
SICOP — Anomaly detection methodology
AreaPublic procurement
Work stream
Demo
HeatSight AI — Demand sensing by geography and SKU
AreaFMCG
Work stream
Essay
Consumer intelligence — Field notes
AreaRetail & FMCG
Work stream
Concept
AI decision engines — Specification
AreaFinancial services
Work stream
Concept
Data governance and maturity
AreaCross-industry
Work stream
Frequently Asked Questions
What people ask about N-AI.
Direct answers to the most common questions we get about the work, services and collaboration models.
Nancy Raquel Rodríguez Ramos is a data and AI strategist, founder of N-AI (Nancy Artificial Intelligence). She has experience designing analytical ecosystems, intelligence architectures and AI-powered solutions across complex business environments. She has worked with the Comptroller General of the Republic of Costa Rica (Contraloría General de la República), marketing agencies and consumer goods (FMCG) companies.
Contact
Let's build intelligent systems.
N-AI engages with executive teams, government innovation leads and category-leading organizations. Reach out with the question you can't answer with the system you have today.
