RL Jones streamlines tariff refund process with AI

RL Jones Customhouse Brokers, a San Diego-based customs brokerage and logistics company, has partnered with Amberd.ai to create an AI decision platform to address data challenges related to the IEEPA tariff refund process.

The Tariff Refund Challenge

Many companies have paid tariffs under the International Emergency Economic Emergency Powers Act (IEEPA). Since a US Supreme Court ruling invalidated those tariffs in February 2026, companies can request a refund on those duties.

“IEEPA tariff refunds represent a significant operational challenge for our business, as the volume and complexity of refund claims has grown substantially with the expansion of tariff actions,” says Eduardo "Lalo" Acosta, VP of RL Jones Group. “Managing these refunds accurately and efficiently is critical not only to our bottom line, but to our clients who depend on us to recover every dollar they are entitled to.”

IEEPA tariff refunds require RL Jones to continuously reconcile CBP downloads, entry data, and customer records under tight deadlines, with refund eligibility and exposure changing as new shipments cross the border. Acosta says the primary challenges stem from the highly variable nature of IEEPA refund claims as each entry can involve multiple tariff codes, overlapping exclusions, and strict documentation requirements that must be reconciled against constantly evolving regulatory guidance.

Previously, RL Jones customs analysts and compliance specialists had to manually review entry documents, cross-reference applicable tariff exclusions, and compile supporting documentation on a case-by-case basis — a labor-intensive process prone to human error. Supervisors also conducted a secondary review before submissions were finalized, meaning a single refund claim could move through multiple hands before it was filed. “Errors or omissions in the filing process can result in denied claims or significant delays, creating both financial exposure and strained client relationships,” says Acosta.

AI Automation

RL Jones deployed the Amberd.ai platform to gain real-time visibility into refund workflows and underlying operations, unifying operational data and documentation into a single AI platform.

“The solution deployed for RL Jones is built on Amberd.ai's core AI platform, which has been purpose-built to address the complexities of modern multi-data source environments,” says Mazda Marvasti, co-founder and Co-CEO of Amberd. “While the foundational architecture is a proven Amberd.ai product, it was configured and refined to align with RL Jones's specific operational workflows and client requirements.”

The two companies worked together to make the solution fit this timely supply chain use case. Marvasti explains, “RL Jones was an integral collaborator in shaping how the platform was deployed, providing critical real-world input on the nuances of IEEPA refund processing that no regulatory document alone could capture. They were a natural fit for this partnership because the scale and complexity of today's trade and logistics environment have simply outpaced what traditional tools and manual processes can sustainably handle, and RL Jones understood that reality. Their willingness to lean into innovation, combined with the escalating cost and risk of conventional approaches, made them exactly the kind of forward-thinking partner with whom Amberd.ai is built to succeed.”

Amberd.ai's platform enables the RL Jones team to ask direct questions about tariff exposure, refund deadlines, and shipment activity and get clear, sourced answers from across their CBP and operational data in a private, governed environment. The platform also automates the most time-consuming elements of the refund workflow — including document ingestion, tariff code classification, exclusion eligibility analysis, and compliance checks — allowing RL Jones to process a greater volume of claims with greater consistency.

The platform is designed to quickly identify who paid specific tariffs, segment customers by tariff exposure, and support refunds when needed; measure eManifest volume, estimate market share percent, and analyze shipment timing and shipper usage; and stay ahead of due dates and detect cycle time and border release issues that can create delays for customers.

However, AI still requires a human in the loop. Acosta confirms, “Human oversight remains an essential part of our process. Our licensed customs brokers and compliance team review and approve all filings before submission, and we make deliberate decisions about when to escalate edge cases for deeper expert review.”

Measurable Improvements

Prior to implementing Amberd.ai, processing a refund claim from initial request to payment could take anywhere from several weeks to over a month, depending on the complexity of the entry and the availability of RL Jones staff, Acosta reports. “With AI-assisted processing, we have reduced that timeline dramatically — in many cases compressing the review and filing cycle to just a matter of days.”

Acosta adds that accuracy has improved measurably, as AI eliminates the inconsistencies that inevitably arise when complex determinations are made manually under time pressure. He has also seen improvements in audit-readiness, since the platform automatically generates a clear, documented trail of how each refund determination was reached — something that previously required significant additional effort to produce.

Acosta says, “We are actively evaluating additional applications of AI across our broader compliance and trade advisory operations, including classification support, country-of-origin analysis, and duty drawback programs. Our experience with IEEPA refunds has given us strong confidence that AI-driven workflows can deliver measurable improvements in both efficiency and accuracy at scale, and we intend to build on that foundation.”

“Tariff volatility has fundamentally changed the calculus for importers and customs brokers,” Marvasti concludes. “What was once a relatively stable compliance environment now requires near-continuous monitoring of executive actions, exclusion windows, and retaliatory measures across multiple trading partners. The velocity of change has simply exceeded what any manual process or traditional software can keep pace with. By the time a human team has analyzed a new regulatory development and updated their workflows accordingly, the landscape may have already shifted again. At Amberd.ai, we believe AI is no longer just a competitive advantage in this environment; it is the only viable mechanism for staying ahead of requirements that are evolving faster than any organization can respond to alone.”

Marvasti adds, “Smart companies like RL Jones are putting in place an AI infrastructure that will drive automation and create competitive advantage as logistics becomes more software-driven and always on.”