NextFin News - In a climate of escalating trade volatility, a group of former Big Tech engineers has emerged from stealth to provide a technological buffer against the rapid-fire policy shifts of the current administration. On February 19, 2026, Amari AI, a startup founded by former Google and LinkedIn veterans, announced it has raised $4.5 million in seed funding to deploy agentic AI systems specifically designed to navigate the complexities of U.S. President Trump’s trade policies. The funding round, co-led by First Round Capital and Pear VC, arrives as customs brokers struggle to keep pace with a regulatory environment defined by sudden tariff hikes and shifting Section 232 investigations.
The startup was co-founded by Sam Basu, a former senior software engineer at Google, and Arushi Vashist, previously of LinkedIn. According to TechCrunch, Basu was inspired to launch the venture after discovering that the $3 trillion U.S. import industry remains heavily reliant on manual paperwork, fax machines, and fragmented spreadsheets. Based in Los Angeles, Amari AI has already secured over 30 customers and facilitated the movement of more than $15 billion in goods. The platform uses custom-trained AI agents to monitor Federal Register updates and social media signals, automatically reclassifying Harmonized Tariff Schedule (HTS) codes and flagging shipments affected by the latest executive orders from U.S. President Trump.
The emergence of such specialized AI tools is a direct response to the "tariff whiplash" that has characterized the first year of the second Trump term. For decades, customs brokerage was a stable, if unglamorous, corner of logistics. However, the current administration’s preference for using tariffs as a primary tool of economic statecraft has turned compliance into a high-stakes game of real-time data processing. When U.S. President Trump announces new duties via social media or sudden press briefings, the lag time between the announcement and the implementation often leaves thousands of containers in transit with uncertain tax liabilities. Amari AI’s agentic architecture is designed to bridge this gap, moving beyond simple data entry to active compliance management.
From an analytical perspective, this trend represents the "verticalization" of generative AI into high-friction regulatory niches. While general-purpose models like GPT-4 have demonstrated broad capabilities, the customs industry requires a level of precision and legal accountability that off-the-shelf solutions cannot provide. Basu noted that the licensing exam for customs brokers has a pass rate of only 10% to 20%, highlighting the extreme complexity of the field. By training models on over one million shipment-related documents, startups are attempting to codify decades of institutional knowledge into digital agents that can work 24/7 without the burnout currently plaguing human compliance officers.
The economic impact of this technological shift is significant. According to industry data, mid-sized brokers using AI-driven automation have reported a 70% reduction in classification research time and a 50% decrease in compliance errors. In an era where U.S. President Trump’s trade team may alter duty rates on short notice, the ability to instantly audit an entire supply chain for exposure is no longer a luxury but a survival requirement. This has created a unique market opportunity: while the broader tech sector has faced a cooling investment climate, enterprise AI startups that solve specific, policy-driven pain points are seeing robust interest from venture capital.
However, the reliance on AI for trade compliance introduces new systemic risks. The customs brokerage industry is notoriously conservative, and for good reason—a single misclassification can result in six-figure fines or the seizure of goods by U.S. Customs and Border Protection. As Vashist and the team at Amari AI scale their operations, the industry must grapple with the "black box" problem of AI decision-making. If an AI agent incorrectly interprets a nuanced policy shift from the White House, the legal liability remains with the human broker. This necessitates a "human-in-the-loop" framework where AI handles the heavy lifting of data parsing while human experts provide the final legal sign-off.
Looking forward, the success of these AI-driven trade tools will likely trigger a broader modernization of the global supply chain. As U.S. President Trump continues to push for a decoupling of key industries from foreign dependencies, the complexity of trade will only increase. We expect to see a surge in "Policy-as-Code" startups that translate political volatility into actionable business intelligence. The move by Basu and Vashist is likely the first of many where Silicon Valley’s elite engineering talent pivots from building consumer apps to building the digital infrastructure required to survive a fragmented and protectionist global economy. In 2026, the most valuable algorithm is no longer the one that predicts what you want to buy, but the one that predicts how much U.S. President Trump’s next tariff will cost you to ship it.
Explore more exclusive insights at nextfin.ai.
