Thirty years of work, across three continents.

Several African central banks faced a duplicated problem: traditional anti-money-laundering (AML) monitoring systems generated thousands of alerts per day, but 95% were false positives. Investigators spent hours closing useless alerts and lost the few relevant cases in the noise. International regulatory pressure (FATF, Basel) demanded ever-higher standards of effectiveness.

Build a transactional monitoring system based on machine learning that (a) drastically reduced false positives without losing critical patterns, (b) was adaptable to different regulatory contexts across African countries, (c) could be adopted by investigation teams without AI background.

Designed and implemented an ML-driven transactional monitoring framework with three key components: multi-level anomaly detection that learns from investigator feedback, supervised pattern recognition transaction scoring, automated compliance process for low-risk cases. The architecture was designed to adapt to different regulatory frameworks without rewriting the core.

40% reduction in false positives. 60% increase in investigation speed. Framework adopted as operational standard. The system is currently active in multiple African jurisdictions with local configurations.

In regulated sectors, AI works when it reduces unnecessary work for experienced operators, not when it promises to replace them. The success was less technical and more organizational: convincing experienced investigators the system was an ally.

The Canary Islands receive approximately 14 million tourists per year, in a regional economy where tourism accounts for over 35% of GDP. Yet the regional government made promotion and resource allocation decisions using fragmented dashboards from different directorates: hotel bookings from one source, airport flows from another, social media sentiment from a third, economic indicators from a fourth. No one had a unified real-time view.

Build an integrated business intelligence ecosystem that aggregated heterogeneous data sources, provided credible seasonal forecasts, and identified opportunities or crises before they became overt.

Designed a unified tourism intelligence architecture with automated ingestion from multiple sources (hotel PMS, flight data APIs, social listening, regional macroeconomic indicators), data normalization and quality layer, predictive algorithms on seasonality and booking patterns, alerting system for sectoral crises. The dashboard was designed to be used both by the regional government and by tourism operators.

15% increase in regional marketing campaign effectiveness. 25% reduction in sectoral crisis response times. The ability to anticipate booking drops 4-6 weeks in advance allowed the regional government to launch corrective campaigns before the season was compromised.

The value of an AI system in a regional government isn't algorithmic sophistication — it's the ability to unify data sources previously read separately. The "single pane of glass" is almost always the first, real, deliverable.

The Principality of Monaco, in line with its tradition of early adoption of strategic technologies, launched a public services transformation program based on AI and blockchain. The challenge wasn't technical — there was no shortage of tech consultants — but governance: how do you adopt technologies that raise questions of transparency, accountability, and citizen rights, in a public administration that must survive changes of government?

Define a governance framework for AI and blockchain adoption in Monegasque public services that balanced innovation capacity, protection of citizen rights, compatibility with European norms (GDPR, emerging AI Act at the time), survival of the political cycle.

Conducted a thorough assessment of existing infrastructure, analyzed priority use cases (digital registries, sovereign identity, citizen services, document management), and designed the governance framework in three layers: high-level principles (transparency, non-discrimination, accountability), operational processes (AI review board, validation processes, periodic audits), technical tools (registry of AI systems, monitoring dashboard).

AI governance framework adopted as internal standard for Monegasque public administration. Monaco positioned as one of the European reference points for responsible adoption of emerging technologies in the public sector. The structure survives the political cycle because it's rooted in operational processes, not statements of intent.

In public institutions, governance isn't a brake on innovation: it's the condition of its survival. Without robust governance, every technological innovation gets canceled or suspended at the first administrative change.