How Tax Authorities Use AI to Obtain Extensive Insights into Taxpayers 


Tax authorities around the world, including several in Africa, are adopting the same Artificial Intelligence (AI) technologies increasingly used by private sector companies, to obtain a 360-degree view of their customers and extensive insights. This is in a concerted effort to the high incidence of tax evasion experienced in many countries. 

According to Dennis Lupambo, Managing Executive at Mint Group Africa,  AI has come a long way in recent years making it easier for companies to aggregate all customer data onto an AI-driven Customer Data Platform (CDP).  

He points out that private sector companies, particularly those in the financial services and retail sectors, have worked tirelessly, and devoted considerable time, effort, and money, to unify and ingest their customer data in an effort to obtain insights into individual customer needs, preferences, attitudes, and behaviors. Often, these insights are spread across various internal and external sources.  

Disparate Data Systems are Proving to be Inefficient  

Generally, the focus is on the customer, optimizing the customer journey, developing new services, and addressing customer issues. Leveraging data also works to the benefit of the business, often helping them to make smarter decisions and differentiate themselves from competitors by building significant, tailored, and valuable solutions.  

The challenge facing companies is that the sources of customer data are often dispersed – held in different departments within the same company, and different subsidiaries across the same group. There is also considerable value to be gleaned from data that are available from different customer-facing touchpoints, ranging from face-to-face or remote, call-center-based interactions to email and social media.   

The Power of a Customer Data Platform  

By drawing on AI’s data matching and analytical power and its tolerance of inaccuracies, CDPs can easily collect data from all existing legacy systems. It can then analyze this data, regardless of the format in which it is presented – and without being overwhelmed by the sheer quantity of data that becomes available.  

This not only eliminates the need for complex data integration or data warehouse projects, it also delivers significant value. Unlike data warehouses that lack analytical capabilities, the AI in CDPs quickly turns raw data into easily accessible and meaningful customer insights. 

How AI is Used to Retain Customers  

Financial sector companies, such as banks, can use these insights to improve their fraud detection, for more accurate credit scoring, and for complying with statutory legislation around Anti-Money-Laundering and Know Your Customer requirements. 

In addition, the information can be used to better segment customers to enable more targeted marketing and promotional campaigns. This helps to increase the customers’ lifetime value and reduce customer churn. 

Lupambo points out that with the bottom 80% of customers bringing in only 20% of business value, the importance of businesses having the ability to identify their most valuable customers cannot be over-emphasized.  

It’s widely estimated that it can cost five times more to attract a new customer than to retain an existing one. In fact, increasing customer retention rates by 5% can translate into a 25 to 95% improvement in profits. 

The Growing Investment in AI-Powered Industries  

The powerful AI-driven CPD technology is being widely used across a host of industry sectors. While financial services – banking and insurance – and retail are probably at the forefront of the CDP trend, many others including the mobile and telecommunications, healthcare, and education industries are also starting to make strides in their move to attain 360-degree views of their clients, customers, students, and patients.  

“Every industry which wants to provide a better service to its customers, or to improve its own productivity and efficacy can benefit from this technology,” says Lupambo. 

This is extending to the government sector with tax authorities, including many in sub-Saharan Africa currently investigating how AI and CDPs can help them boost their tax collections and slow or even reverse the trend of shrinking tax revenues.  

Lupambo says that effective tax collection has become an increasingly urgent challenge for almost every revenue authority on the continent. 

Increasing the Efficiency of Tax Collections   

A study conducted by the UN Economic Commission for Africa (UNECA) found that improved tax governance and efficiency could deliver as much as US$110 billion in new tax revenue between 2020 and 2025 – more than double the US$51.8 billion Africa received in official development assistance in 2018. 

“In much the same way as private sector companies use CDPs to obtain comprehensive insights into their customers, tax authorities can flesh out their partial view of taxpayers by following the digital footprints we all leave, whether we realize it or not,” Lupambo says.  

“Even if an individual attempts to go off the digital grid by, for example, only using cash and avoiding social media, they will leave a trail somewhere.” 

Purposeful Data Processing  

Legislation such as the Protection of Private Information (POPI) Act doesn’t protect such information from the government mandate to obtain information on citizens and residents for tax purposes. However, tax authorities don’t necessarily have to access information held by banks, for example. 

They could obtain a great deal of data simply by trawling the databases in many different government departments, from Customs and Excise to the Deeds Office where all fixed property is registered, or even vehicle licensing departments.  


“By tapping these sources of information and combining it with the power of AIMachine Learning,  and powerful technology, the tax authority will be able to segment and predict which taxpayers to investigate. Indeed, the need for cumbersome lifestyle audits may also be reduced as much of the anomalies in terms of declared income and undeclared assets could be readily laid bare,” Lupambo concludes. 

Recent Blogs