Any customs system not built on Large language model (LLM) will Disappear,” warns ASYCUDA founder Jean Gurunlian

Chairperson of Webb Fontaine, and founder/architect of the ASYCUDA system, Jean Gurunlian, has issued a stark warning on the future of Customs technology at the WCO Technology Conference 2026 in Abu Dhabi, stating that legacy Customs systems are no longer fit for purpose in a world shaped by large language models (LLMs).

Gurunlian said “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.”

Gurunlian, who designed ASYCUDA and oversaw its deployment in more than 100 countries, emphasized that the pace and nature of regulatory change have fundamentally shifted. Traditional Customs systems, often reliant on static rules, manual updates, and lengthy development cycles, can no longer cope with today’s trade and policy environment.

“Customs systems that cannot adapt to changing laws, regulations, or operational requirements within very short timeframes simply will not survive anymore,” he said. “If a system needs years to adjust, it is already too late.”

According to Gurunlian, the rise of LLMs has exposed the structural weakness of systems that depend on predefined logic rather than continuous learning and adaptation. He noted that many Customs platforms still require months, or even years, to incorporate legislative changes, tariff updates, or new non-tariff measures.

“Tariffs and non-tariff barriers have increasingly become political weapons,” Gurunlian said. “They can change overnight, sometimes without warning. With LLM-enabled systems, those changes can be interpreted, applied, and operationalized in seconds.”

For Gurunlian, adaptability is no longer a feature but a prerequisite. He underscored that Customs technology must now be designed around continuous improvement, contextual understanding, and rapid learning, capabilities that only AI-driven architectures can provide.

“The systems, including those that I created, are bound to become obsolete,” Gurunlian acknowledged. “If a system cannot be improved in production, it should not be deployed.”

He added that governments and Customs administrations face a critical decision point. Continuing to invest in static, rule-based systems risks locking institutions into technology that cannot respond to geopolitical shocks, regulatory volatility, or the increasing complexity of global trade.

“LLMs change the nature of systems themselves,” Gurunlian concluded. “This is not about adding AI on top of existing platforms. It is about rethinking Customs systems from the ground up.”