In a few months since its deployment, the Ghana Revenue Authority’s (GRA) artificial intelligence system, Publican AI, has shifted the dynamics of trade compliance at the nation’s ports. Despite initial resistance from freight forwarders and traders, the technology has reportedly generated an additional GH¢1 billion in customs revenue for April 2026 alone, according to recent data released by the Authority .
The system, which went live at the Tema Port between February & March represents the government’s most aggressive push to date to dismantle deep-rooted customs fraud and close valuation leakages that have historically cost the state billions . Officials now report that the “hybrid model” is validating shipments with a speed and accuracy impossible for manual processes.
The History of Publican AI
Publican AI did not emerge in a vacuum but is the culmination of years of policy planning and technological procurement aimed at solving a specific fiscal crisis. For years, the GRA struggled with what Commissioner-General Anthony Sarpong described as “human-dependent processes that leave room for discretion and manipulation”. Between 2020 and 2024, the government estimated it lost billions in potential customs revenue due to rampant under-declaration, misclassification of goods, and outright evasion.
The road to automation formally began in late 2025 when the GRA, backed by the Ministry of Finance, secured parliamentary approval for the “Publican Trade Solution.” Unlike the previous Integrated Customs Management System (ICUMS)—which faced similar resistance during its rollout in 2020—Publican AI was designed specifically as a machine-learning enforcement tool rather than just a processing database .
How the System Works
Speaking during a media briefing in late April, Commissioner-General Sarpong explained that Publican AI functions as a decision-support tool that interfaces with global trade databases. It reviews three critical pressure points: the country of origin, the classification code of the product, and the valuation price .
Previously, a customs officer might take up to two hours to manually verify a single document. The AI currently performs this function in roughly five minutes, flagging approximately 25 per cent of all declarations for further human review while automatically clearing the rest .
“We are not taking away human oversight,” Sarpong assured stakeholders during a meeting with the Association of Ghana Industries (AGI). “The system provides faster and more reliable data, but the final decision still rests with the officer.”
Revenue Impact
The financial results have been immediate and substantial. In early April, Sarpong revealed that the system was generating an average of $3 million (approx. GH¢41 million) in extra daily revenue—a figure that translates to nearly $1 billion monthly if fully optimized .
By June, the GRA was able to quantify the success more concretely, linking an additional GH¢1 billion in April revenue directly to the AI’s deployment. This suggests that the system is recovering funds that were previously lost to practices like under-invoicing, where importers would declare lower values for luxury cars, machinery, and consumer goods to evade the full duty rate .
“Our aspiration for a reset and transformed economy is not attainable if we fail to mobilise the needed domestic revenue,” Sarpong stated at the 10th Ghana CEO Summit, emphasizing that this digital reform is key to reducing Ghana’s reliance on external financing .
Industry Divisions
However, the path to digital efficiency has been turbulent. In the weeks following the rollout, tensions escalated dramatically. The Traders Advocacy Group Ghana (TAGG) and the Ghana Union of Traders Association (GUTA) led protests, with some groups wearing red armbands and abandoning goods at the ports in what they described as a sit-down strike .
Protestors argue that the AI has removed necessary flexibility, leading to inflated and unsustainable duty assessments. The Coalition of Concerned Exporters, Importers, and Freight Forwarders alleged that the system forces officers to accept AI-generated values even when they appear inconsistent with the condition of the actual goods .
While retail traders have protested, major manufacturing and industrial bodies have largely endorsed the reform. The Association of Ghana Industries (AGI) stated that its members—who constitute about 80 per cent of the manufacturing sector—are satisfied that the system promotes fairness.
“We are not against the system but needed to understand the system and get better clarifications,” said Dr. Kofi Nsiah-Poku, President of the AGI, following a closed-door meeting with the GRA. He noted that the AI helps ensure that legitimate manufacturers are not undercut by smugglers or under-declarers .
Similarly, the Importers and Exporters Association of Ghana has backed the rollout, arguing that the system ends the arbitrary discretion that often punished small-scale importers while allowing larger operators to evade taxes .
The Future
As Ghana navigates this transition, the government remains committed to expanding the system beyond Tema to the Takoradi Port and all land borders. While the “friction” of algorithmic enforcement continues to cause congestion and slower clearance times for some high-risk shipments, officials argue this is a necessary trade-off for fiscal integrity .
For now, the numbers are compelling. With GH¢1 billion added to state coffers in a single month, Publican AI has moved from a controversial experiment to a central pillar of Ghana’s revenue mobilization strategy.