The Ghana Revenue Authority (GRA) says the roll out of an Artificial Intelligence (AI)‑powered valuation and risk‑assessment system is to help curb high‑risk import practices at Ghana’s ports such as misclassification, undervaluation and incorrect declaration of country of origin.
It is also to level the playing field for compliant businesses and boost domestic revenue mobilisation.
Speaking at a forum organised by the Association of Ghana Industries (AGI), the Commissioner‑General of the GRA, Mr. Anthony Kwasi Sarpong, said the technology is intended to improve fairness at the ports, strengthen competitiveness among businesses, and protect government revenue.
Mr. Sarpong said that upon assuming office, management conducted a review of five years of import and foreign exchange data, which exposed significant inconsistencies between imports declared and payments made abroad.
“Our analysis showed that nearly US$83 billion left the country over a five‑year period to pay for imports, yet only about US$31 billion worth of goods was reflected in customs declarations,” he disclosed. “That gap raised serious red flags.”
According to him, the discrepancy suggested either large‑scale under‑declaration of imports or unauthorised financial outflows—both of which undermine the tax system and the broader economy.
Based on conservative estimates, the GRA concluded that the state could be losing about US$11 billion annually in potential tax revenue.
Mr. Sarpong stressed that the issue was beyond revenue losses and directly affected business sustainability.
“When one business declares accurately, pays the correct duties and taxes, and another does not, the result is price distortion,” he said.
“The compliant business becomes uncompetitive and is gradually pushed out of the market.”
He noted that this situation eroded confidence in the system and discouraged honest businesses, ultimately slowing overall economic growth.
The Commissioner‑General explained that previous customs processes were heavily dependent on manual intervention, creating room for excessive discretion and, in some cases, collusion between officers, clearing agents and importers.
“We decided to reduce human discretion by introducing a neutral system—one that does not know business A or business B,” he said. “That is where AI becomes critical.”
The AI system focuses on four key risk areas country of origin and port of loading, tariff classification, customs valuation and non‑duty related risks, including expired drugs, prohibited substances and other dangerous goods.
Under the new process, once an import declaration is submitted, the AI automatically analyses it—cross‑checking global pricing data, historical records and country‑of‑origin details—and generates a risk report for the customs officer within minutes.
“For valuation, the system identifies the lowest and highest prices at which similar goods have been imported elsewhere,” Mr. Sarpong explained. “If a declared value falls far outside that range, the system flags it for review.”
During the pilot phase, the AI tool was applied silently across all import declarations.
“Results showed that about 75 percent of declarations were consistent with global benchmarks,” he said. “However, 25 percent were flagged for potential misclassification, undervaluation or other risk factors.”
He emphasised that the AI was not to replace customs officers but to support them.
“The system does the heavy analytical work. The officer reviews the findings and determines whether additional checks or documentation are required,” he said.
Mr. Sarpong assured importers and clearing agents that all existing appeal mechanisms remain in place.
“If an importer disagrees with an assessment, the case can still be reviewed by senior customs officers, just as it has always been,” he said, adding that the AI introduction has actually added another layer of review for greater transparency.
Responding to industry concerns, the GRA boss clarified that the AI system had not introduced any additional fees or levies.
“This reform is fully funded by government,” he said. “Where duties increase, it is only because the correct value or classification has been applied—not because of a new charge.”
Mr. Sarpong concluded that the ultimate goal of the initiative is fairness.
“When everyone is assessed using the same baseline, businesses compete fairly, trade is facilitated, and government revenue is protected,” he said.
The session ended with a live demonstration of the AI system by GRA technical officials.
By: Christian Akorlie / businesspostonline
