Host: Annik Sobing
Guests: Jennifer Varney (Volvo Group), Penny Chen (PAX)
Recorded at: ICPA Conference, San Antonio, TX
Published: March 2026
Length: ~25 minutes
Presented by: Global Training Center

AI Meets Trade Compliance: From Auto Supply Chains to AI

Live from ICPA San Antonio, Annik sits down with Jennifer from Volvo Group and Penny from PAX for an all‑women, International Women’s Day‑timed conversation about how AI is actually being used in trade compliance today—far beyond the buzzwords. They explore the reality of AI inside a massively complex automotive supply chain, how duty drawback is being reimagined with AI, and what trade teams should think about before buying or building any tools.

What You’ll Learn in This Episode

Session highlights from ICPA
- Jennifer: Practical implementation of AI to support customs clearance at the enterprise level—how one company uses AI to survive an “ever‑changing and incredibly volatile” trade landscape.
- Penny: A “beginner‑friendly” intro to general AI tools, how large language models work, and how trade compliance leaders can evaluate AI quality and fit.
The automotive reality: 1,000+ policy changes and thousands of parts
- In just the last year, there have been 1,000+ trade policy changes worldwide, affecting about 5 trillion dollars in spend.
- Most of the real impact comes from trade barrier changes, not facilitation measures.
- A single vehicle can have 2,000–3,000 parts sourced from thousands of suppliers globally, some in‑house, some external.
- New demands around Section 232 (steel/aluminum/copper), forced labor, EUDR, connected vehicle rules, dual‑use, etc. mean OEMs must know their supply base down to raw material origin and processing, sometimes 5–6 tiers deep.
Why human-only workflows can’t keep up
- Many tier‑1 suppliers don’t even have the data OEMs now must report, or consider it proprietary.
- Trade teams are drowning in documentation, entry creation, and ever‑changing regulatory demands—falling behind risks blocked shipments and massive cost.
- Jennifer’s view: AI is less about replacing people and more about augmenting limited resources before they’re “buried under all of the legislative changes.”
Where AI fits in (and where it doesn’t)
- Example use case: consolidating multiple documents (PO, invoice, BL, shipping manifest) to build a single 7501—AI reads different formats, extracts the right fields, and populates data so humans review instead of retyping.
- Penny’s rule of thumb: if it’s a task you’d happily delegate to an intern, it’s a candidate for automation or semi‑automation.
- AI frees people to focus on high‑value work: audits, wider coverage (5% → 99%), forecasting regulatory changes, and adjusting systems/processes for what’s coming next.
Starting your AI journey: practical adoption path
- Step 1: Use free or existing tools (e.g., Microsoft Copilot) for summaries, data cleaning, and simple tasks.
- Step 2: When needs get more complex, consider specialized AI tools (like PAX’s AI‑powered duty drawback service), but pair them with solid ROI analysis: cost vs. time savings vs. recovered dollars.
- Step 3: For large enterprises, begin with defining pain points and a data strategy:
  - Where do you spend the most time?
  - Which activity is eating 90% of your bandwidth?
  - What data will go into AI, and what exactly do you want back out?
Overcoming fear and building buy‑in
- Penny’s take: curiosity is your best ally—if you don’t know how to use AI, start by asking AI how to use AI.
- ...