By 
LLMs Have Rendered Traditional Customs Platforms Obsolete – Gurunlian
Jean Gurunlian, Chairperson of Webb Fontaine and the founder and chief architect of the globally deployed ASYCUDA system, has warned that legacy Customs technologies are no longer viable in a world increasingly shaped by large language models (LLMs).
RELATED: Nigeria Customs Service launches B’Odogwu: Unified management system to enhance trade efficiency
Speaking at the World Customs Organization (WCO) Technology Conference 2026 in Abu Dhabi, Gurunlian declared that Customs systems not designed around LLM-driven architectures are already obsolete.
“No Customs system that has not been built on LLM will survive. The first real outcome of LLMs is that they have made all existing Customs systems obsolete,” he said.
Why Traditional Customs Systems Can No Longer Keep Pace
Leveraging decades of global expertise from deploying ASYCUDA in over 100 countries, Gurunlian highlighted how today’s regulatory and trade environment has outpaced traditional, rule-based Customs platforms. The increasing pace, complexity, and volatility of global commerce now demand more advanced and agile technological solutions.
He noted that most existing systems depend on static logic, manual configuration, and lengthy development cycles. This makes them incapable of responding to rapidly changing laws, tariffs, and operational requirements.
“Customs systems that cannot adapt to changing laws, regulations, or operational requirements within very short timeframes simply will not survive anymore. If a system needs years to adjust, it is already too late,” he warned.
LLMs Expose Structural Weakness in Rule-Based Platforms
According to Gurunlian, the advent of LLMs has exposed critical limitations in traditional, rule-based Customs systems. These legacy platforms lack the capacity for continuous learning and contextual analysis. All of these are now essential in modern, dynamic trade environments.
He noted that many existing systems still require months or even years to implement legislative updates, tariff shifts, or new non-tariff measures. This prolonged adaptation period leaves customs administrations vulnerable to compliance gaps and operational inefficiencies.
“Tariffs and non-tariff barriers have increasingly become political weapons. They can change overnight, sometimes without warning. With LLM-enabled systems, those changes can be interpreted, applied, and operationalised in seconds,” he said.
Adaptability Is No Longer Optional
Gurunlian stressed that adaptability is no longer a competitive advantage but a basic requirement for Customs administrations operating in today’s geopolitical and economic climate.
He argued that future-ready Customs systems must be built for continuous improvement, rapid learning, and contextual decision-making. These are capabilities inherent to AI-native and LLM-driven architectures.
Jean Gurunlian
“The systems, including those that I created, are bound to become obsolete. If a system cannot be improved in production, it should not be deployed,” Gurunlian acknowledged.
A Critical Decision Point for Governments and Customs Authorities
Gurunlian cautioned governments and Customs agencies against continued investment in static, rule-based systems. He warned that such choices risk locking institutions into technologies unable to withstand regulatory volatility. This is in adition to geopolitical shocks, or the growing complexity of global trade.
He emphasised that the transition required is not about retrofitting AI onto legacy platforms.It is about fundamentally rethinking how Customs systems are designed.
“LLMs change the nature of systems themselves. This is not about adding AI on top of existing platforms. It is about rethinking Customs systems from the ground up,” he concluded.
