KRA to deploy AI system for analysis of business registration data

The Kenya Revenue Authority is moving to integrate an artificial intelligence system designed to scrutinize data from the Business Registration Service. This technical shift aims to provide the tax authority with clearer oversight of corporate structures, identifying the specific individuals who own and direct companies operating within the country.

By linking directly with the Business Registration Service, the system allows the taxman to verify whether businesses are declaring and paying the correct taxes based on their registered activities. This level of oversight is expected to have particular relevance for the construction and infrastructure sectors, where complex joint ventures and multi-layered subcontracting arrangements are common.

The implementation follows a broader digital transformation strategy at Times Tower, which includes the automated validation of tax returns. Under this new framework, the authority can flag inconsistencies between reported income and third-party data sources in real-time. For firms in the built environment, this means that every declared expense must be supported by verifiable digital records, largely through the Electronic Tax Invoice Management System.

Construction firms often manage diverse supply chains, and the new AI capabilities are designed to identify suspicious patterns across these networks. The system will employ advanced visualization features, such as social network analysis and geospatial mapping, to help investigators interpret complex datasets. This enables the tax authority to build detailed taxpayer profiles and identify potential fraud or under-reporting long before a manual audit would typically be triggered.

From January 2026, the validation engine began cross-referencing income and expenses in both individual and corporate tax returns against TIMS and eTIMS invoices, withholding tax amounts, and customs import records. Any expense claim filed by a business that lacks a corresponding digital invoice in the system is now subject to immediate rejection, effectively increasing the tax liability for non-compliant entities.

The move is part of the Tax Administration 3.0 blueprint, which seeks to reduce human error and eliminate loopholes in revenue collection. While the authority maintains that these tools will foster trust and transparency, the rollout has prompted discussions regarding the intersection of state surveillance and financial privacy. For directors of construction companies, the focus remains on ensuring that all procurement and payroll data aligns strictly with the digital trails monitored by the taxman.

As the state intensifies its drive to expand the tax base, the use of algorithms to perform what was once manual auditing marks a shift toward a technology-driven system manager model. The primary objective remains the identification of high-income entities and professionals who may be operating outside the formal tax net despite significant participation in the economy.