Jean Gurunlian, Chairperson of Webb Fontaine and founder of the ASYCUDA customs system, has warned that traditional customs technology is approaching extinction unless it is rebuilt around large language models and artificial intelligence.
He delivered the message at the World Customs Organization (WCO) Technology Conference 2026 in Abu Dhabi, where policymakers, technology leaders, and customs authorities gathered to discuss the future of border management and global trade systems.
According to Gurunlian, customs platforms that rely on static rules, manual updates, and long development cycles can no longer operate effectively in today’s fast-changing trade environment.
“No Customs system that has not been built on large language models will survive,” he said. “The first real outcome of LLMs is that they have made existing systems obsolete.”
Gurunlian designed ASYCUDA more than three decades ago and oversaw its deployment in over 100 countries, including many across Africa. The platform became one of the most widely adopted digital customs systems in the developing world, helping governments automate revenue collection and cargo clearance.
However, he now argues that the pace of regulatory and political change has outgrown the capabilities of legacy technology.
Modern trade policies shift rapidly, driven by geopolitics, sanctions, security concerns, and economic competition. Tariffs and non-tariff barriers often change with little notice, placing heavy pressure on customs administrations to respond immediately.
Yet many existing systems still require months, and sometimes years, to integrate new regulations, tariff schedules, or compliance rules.
“Customs systems that cannot adapt to changing laws or operational requirements within very short timeframes will not survive anymore,” Gurunlian said. “If a system needs years to adjust, it is already too late.”
He explained that large language models introduce a different technological foundation. Instead of depending on predefined logic and slow software updates, AI-driven platforms can interpret new regulations, translate them into operational rules, and apply them in near real time.
“Tariffs and trade restrictions have become political weapons,” he added. “They can change overnight. With LLM-enabled systems, those changes can be understood and operationalised in seconds.”
Gurunlian described adaptability as the new baseline for public infrastructure, not an optional feature. He stressed that continuous learning, contextual understanding, and rapid system improvement must now sit at the core of customs technology.
In a notable admission, he acknowledged that even ASYCUDA belongs to a technological generation that is nearing its limits.
“The systems, including those that I created, are bound to become obsolete,” he said. “If a system cannot be improved in production, it should not be deployed.”
His remarks place renewed attention on the choices facing governments, particularly in emerging markets and African economies where ASYCUDA remains widely used.
Customs operations play a central role in national revenue, border security, trade facilitation, and investor confidence. Slow clearance processes increase business costs, weaken supply chains, and reduce competitiveness in global markets.
With the African Continental Free Trade Area (AfCFTA) aiming to boost intra-African trade, customs efficiency has also become a strategic economic priority for the continent.
Technology analysts at the conference noted that AI-native customs platforms could shorten clearance times, improve risk profiling, detect fraud more effectively, and support real-time compliance with complex international trade rules.
Gurunlian cautioned, however, that meaningful progress requires more than adding artificial intelligence to existing platforms.
“LLMs change the nature of systems themselves,” he said. “This is not about placing AI on top of old technology. It is about rethinking customs systems from the ground up.”
As global trade becomes more digital, politically sensitive, and interconnected, his warning signals a broader shift in how governments approach digital infrastructure.
For many countries, the next generation of customs systems may determine not only how goods move across borders, but also how effectively they compete in an increasingly technology-driven global economy.
