AI in cross-border tax administration

The power of machine learning and advanced analytics is not about replacing human judgment, it is about amplifying it. Supervised models can score risk across millions of records so auditors focus on the highest-impact cases.

Unsupervised learning can surface unusual clusters of behaviour that traditional rules miss. Time-series forecasting helps revenue managers anticipate seasonal cash flows and inform policy choices.

Natural language processing can convert unstructured documents, contracts and shipping manifests into searchable tax intelligence. In short, artificial intelligence (AI) and analytics turn scattered data into a continuous decision support system for the tax office.

These capabilities are already shaping outcomes across the region. Tax administrations that combine richer data feeds with analytic platforms report improved detection of fraud, faster case resolution and more precise targeting of compliance interventions.

The OECD’s recent work on tax administration shows that many countries are deploying application programming interfaces (APIs) and automated data links, alongside stronger data governance and analytics.

Across East Africa, countries are pursuing complementary digital approaches that we can learn from. Tanzania’s use of electronic tax stamps has been credited with substantial gains in excise administration and supply-chain visibility. IPP Media Uganda has expanded electronic fiscal and invoicing systems as part of a wider push to widen the tax base and reduce leakage.

PwC Rwanda has implemented mandatory electronic invoicing and real-time invoice reporting that strengthens VAT invoice capture and has become a regional case study for e-invoicing design and outcomes.

Meanwhile, the East African Community is advancing digital integration and customs interoperability to ease cross-border trade and improve the flow of trade data across partners.

These national and regional efforts offer an important lesson: digital tools produce the greatest public value when they combine automated data flows, strong analytics and legal and operational frameworks.

In Kenya, recent initiatives to modernise border processing are designed to reduce clearance times and make trade reporting more seamless, part of a wider move toward interoperable systems that support both trade facilitation and revenue assurance.

But with great data power comes great responsibility. As tax administrations become more data-rich, legitimate public concern about privacy, surveillance and misuse of information increases.

We must, therefore treat data protection and explainability as core design principles. This means: strong legal frameworks that specify purpose, scope and retention; APIs that restrict access by purpose and jurisdiction; encryption and role-based access controls; model-explainability tools so decisions can be audited; and human-in-the-loop review for automated high-stakes actions.

Building public trust also requires transparency about how algorithms are developed and used.

Explainable risk scores, clear redress mechanisms, regular fairness audits and independent oversight help to ensure AI tools support equitable enforcement rather than opaque profiling.

Where cross-border sharing is necessary, we should adopt process-based interfaces (narrow, logged API calls with jurisdictional access limits) and operate them within bilateral or regional legal agreements that protect citizens’ privacy.

When designed and governed on the pillars of privacy, accountability and responsible innovation, AI-powered revenue intelligence can make tax systems smarter and more trusted. It can shift effort away from manual checks and low-value admin to higher-value services: faster refunds for taxpayers, lighter touch compliance for small honest firms, and focus on sophisticated evasion and fraud.

Kenya’s ambition should therefore be twofold: to accelerate the adoption of analytics, machine learning and secure data engineering across the Kenya Revenue Authority’s systems; and to deepen regional cooperation on standards and interoperable exchanges so that cross-border trade works for citizens, commerce and the public purse.

By combining intelligent tools, rigorous governance and regional collaboration, we can build a revenue environment that is transparent, predictable and fair, one that supports Kenya’s broader economic ambitions while protecting the rights of taxpayers.