The Federal Inland Income Service’s (FIRS) TaxPro Max platform, launched to modernise and allow e-tax submitting and funds by firms, represents a serious step in Nigeria’s digital tax assortment, but it surely lacks the superior predictive capabilities wanted to detect subtle tax fraud and income manipulation, a knowledge scientist, Mr. Emeka Atuma, has stated.
Atuma, an unbiased advisor and coach in Knowledge Science, made this assertion on Friday on the tenth Blakey’s Nationwide Tax Convention held in Lagos, the place he introduced a paper titled “Knowledge and Synthetic Intelligence in Administering Tax in Nigeria.”
The Principal Accomplice and CEO of Administration Consulting and Enterprise Advisory Companies said that whereas TaxPro Max can enhance compliance by means of digital submitting, it can not precisely establish tax avoidance schemes and extra advanced issues.
Digital submitting with out predictive intelligence
Based on Atuma, TaxPro Max largely features as an digital submitting and cost platform, with little analytical depth to evaluate the credibility of returns submitted by taxpayers.
“TaxPro Max is a begin, but it surely doesn’t have the predictive analytics to find out whether or not a submitting is fraudulent or whether or not a taxpayer is intentionally manipulating the system,” stated the Knowledge Scientist.
He defined that superior Synthetic Intelligence (AI) programs can analyse historic knowledge, detect uncommon patterns and flag suspicious filings—akin to firms persistently maintaining income just under tax thresholds to cut back liabilities—however such intelligence is at the moment lacking in FIRS’ TaxPromax.
Atuma’s explanations have been in response to a member of the viewers who complained about points together with his e-filing the place he was wrongly assessed and an inaccurate tax quantity was imposed on his agency and was by no means reviewed regardless of his protests.
“All I used to be informed is go and pay,” the participant within the convention claimed.
Why AI issues for tax administration
Atuma famous that AI combines knowledge science, machine studying and deep studying fashions to analyse massive datasets and establish compliance gaps. As soon as dependable knowledge is accessible, AI can distinguish between compliant taxpayers and people at excessive danger of evasion, permitting tax authorities to focus audits the place they matter most.
Nonetheless, he pressured that AI doesn’t work in isolation. “AI can solely operate the place high quality knowledge exists,” he stated, including that poor knowledge integration throughout authorities companies might undermine tax assortment effectivity.
Knowledge gaps and infrastructure challenges
The knowledgeable recognized fragmented taxpayer knowledge as one in all FIRS’ largest constraints. Data on the identical people and companies is unfold throughout a number of companies, together with the Company Affairs Fee, Nationwide Id Administration Fee and the Central Bank of Nigeria.
He insisted till these units of information centres are aggregated, cleaned up and standardised, they may not be adequate to work with in superior AI analytics.
He additionally cited infrastructure deficits—unreliable energy provide, restricted bandwidth and outdated legacy programs—as limitations to deploying superior analytics at scale.
Prolong TaxPro Max, don’t change it
Slightly than scrapping TaxPro Max, Atuma really useful extending the platform by embedding machine studying algorithms behind it.
This might permit the system to analyse saved tax returns, flag anomalies and help risk-based audits with out changing human judgement.
Such fashions, he stated, can cluster taxpayers with comparable behaviour and spotlight filings that deviate sharply from established patterns for additional investigation.
Nonetheless, he pressured that privateness legal guidelines forestall the federal government from straight accessing residents’ bank accounts at the same time as he famous that they’ve entry to bank accounts of people and company organisations.
He’s of the view that AI programs ought to as an alternative depend on oblique verification, notifying taxpayers of discrepancies and requesting clarification by means of due course of.
Casual economic system nonetheless a blind spot
Atuma added that even essentially the most superior AI system would battle to seize Nigeria’s massive casual economic system with out broader knowledge assortment mechanisms. Many economically lively people function exterior the banking system, limiting visibility for tax authorities.
To deal with this, he recommended increasing knowledge seize factors past banks to areas akin to SIM registration, healthcare entry and different on a regular basis providers that require identification verification.
“AI can not work exterior obtainable knowledge,” he stated, including that telecoms knowledge might present a essential entry level for figuring out economically lively people with out BVN, NIN, TIN or bank accounts.
Be First to Comment