In the space of a single week in July, the U.S. government published new “Brazil 301” and “Forced Labor 301” tariff schedules. By the time my team finished parsing the actions that covered the top 60 U.S. trading partners, representing 99.4% of U.S. imports, mapping every carve-out, and cross-referencing the exemption appendices, the update required generating 11,000 distinct compliance rules. 

Read also: Tariff Refunds, Economic Uncertainty, and What’s Ahead

Not 11,000 line items. 11,000 rules, each governing a specific commodity from a specific country, imported on a specific date, to determine the chapter 99 codes and associated duty impact once the product gets admitted to the zone.

Every FTZ program in the country was required to absorb that same update. The ones running systems that handle this automatically didn’t notice. The ones doing it by hand are still catching up.  

This is the environment FTZ programs are operating in right now. And new research from Dimensional Research, based on a survey of 301 VP and C-level executives in trade compliance, supply chain, and operations, shows that most of the software managing these programs was not built for this level of change.

The Numbers Are Specific Enough to Be Uncomfortable

Only 20% of respondents say their FTZ software automatically updates when tariff schedules change. That means 80% of organizations managing active zones right now are updating tariff rules by hand, zone by zone, system by system, in a regulatory environment where the schedules can shift without notice and the window between a change and a compliance failure is not forgiving. Ninety-seven percent are running their FTZ operations across three or more disconnected software applications. Spreadsheets, ERP modules, dedicated FTZ tools, custom-built workarounds. Each system is holding a piece of the compliance picture, but none are holding all of it.

I have spent 20 years in this space, and what surprises me about these numbers is not that they’re high, it’s that they’re consistent. The patchwork architecture these programs are running on was not the result of bad decisions. It was assembled over time, zone by zone, as programs grew, as requirements changed, as new regulatory obligations got bolted onto existing infrastructure. The workarounds accumulated but softwaree didn’t adapt. For a long time, the pace of change in the trade environment was slow enough that the gap between what the software could do and what the programs needed didn’t create an acute crisis. It just created friction.

That friction has become something else entirely. When tariff schedules change as frequently as they have over the past two years, the reconciliation lag in a manual update cycle isn’t a minor inefficiency. It causes entries to get rejected leaving importers unable to file. Where penalty exposure accumulates across active shipments. Where a compliance team or finance is working off duty calculations that stopped reflecting reality three weeks ago and won’t know it until the next reconciliation cycle closes.

What the Report Data Actually Shows

The report data in the Dimensional Research survey puts a number to what that friction costs. 67% of respondents experienced FTZ audits with negative findings in the last 11 months. Not over five years, not historically — in the last 11 months. When they were asked what caused those findings, FTZ management software ranked as the leading cause. Not process failures or staffing. Software.

The consequences are severe. Among the executives who experienced negative audit findings: 34% received a minor financial penalty, 27% a major financial penalty, 25% faced operational sanctions, and 23% had to close an FTZ entirely. Site deactivation is worth pausing on. It isn’t a corrective action that resolves after remediation. Deactivating means surrendering every financial benefit the program was generating — duty deferral, cash flow timing, working capital advantages — and CBP won’t approve the deactivation until all merchandise in zone status has been cleared out, which generally means paying duty on inventory that had been sitting deferred. Reactivation is possible, but it happens under the scrutiny that follows an adverse finding.

The timing matters because FTZ programs are not contracting. They’re growing. The same survey found that 98% of companies using FTZs are now running multiple zones, and five of the top seven business changes executives report making in response to tariff volatility involve FTZ utilization. 

Companies are leaning into FTZs precisely because the duty economics make sense — deferral, reduction, elimination — and the programs are expanding to capture more of those advantages. The problem is that the software managing the expanded programs is largely the same software that managed the existing programs, and it wasn’t keeping up with those either.

AI Won’t Fix a Foundation That Isn’t There

Somewhere in the middle of this, 99% of the executives in the survey said they are using or planning to use AI to manage FTZ operations. I understand the appeal. The top intended use cases are compliance checks, document ingestion, and product classification — exactly the tasks that are most painful to do manually and most prone to error under time pressure. But there’s a sequencing question here that I think gets glossed over in most conversations about AI in trade compliance.

Unfortunately for trade, the government does not provide a single exhaustive data feed of all HTS codes and tie them to all applicable additional tariffs and other related information. CSMS alerts — the text-based notifications Customs uses to announce changes — are dense, technical, and full of carve-outs that only apply under specific combinations of origin country, product classification, import date, and export date. The appendices that cover the exceptions come as PDFs. 

None of it arrives in a format that plugs cleanly into a system. What makes the 11,000-rule update possible at all is a layer of infrastructure that monitors these sources continuously, parses them, and translates them into system-enforced rules before a shipment receipt ever arrives. AI applied to that process, when the underlying data is clean and structured and current, can meaningfully accelerate classification, catch anomalies, and reduce manual review time.

AI applied to a program where tariff rules are updated manually, where compliance documentation gets assembled before each audit rather than generated on demand, where inventory data consolidates on a reporting schedule rather than in real time — that’s a different situation. It helps process the gap faster, it doesn’t fix the gap.

The programs that will actually benefit from AI investment in FTZ operations are the ones that have already solved the foundation: automatic tariff updates, an intelligent tariff engine that helps with the chapter 99 assignment, real-time inventory visibility per zone, on-demand audit documentation. Not because those things are prerequisites in some theoretical sense, but because without them, the data AI is working with isn’t reliable enough to produce reliable outputs. The sequencing isn’t a technology preference. It’s an accuracy question.

The Platform That Got the Job Done in 2019 Is in a Different Environment Now

What I used to hear most often from compliance leaders evaluating their current platforms was some version of: “It’s not perfect, but it gets the job done.” That assessment was probably accurate when the rate of change was more manageable. The evaluation criteria that mattered when most of these platforms were selected — compliance reporting, documentation accuracy, CBP filing capability — are still relevant. But the environment has added requirements those platforms weren’t designed to meet. Automatic additional tariff assignment and calculation updates across all zones. Regulatory rule sets that can absorb an 11,000-rule update in a single push rather than a manual cycle. Audit documentation that’s available on demand rather than assembled over weeks. The platform that got the job done in 2019 is operating in a different trade environment now.

The Dimensional Research survey ends with a conclusion I’d put differently than a market research firm would, but I don’t disagree with it: FTZs are a critical strategy for managing tariff risk, but the software being used isn’t integrated and lacks key functionality. What I’d add is that the gap between FTZ program growth and FTZ software capability has probably never been wider, and tariff volatility is making the cost of that gap visible in ways it wasn’t before. The compliance teams working hardest right now are often the ones whose organizations have invested the most in FTZ strategy. The strategy is sound. The infrastructure running it needs to catch up.

Author Bio

Joshua Guy is VP of Global Trade Compliance at QAD, where he leads the company’s FTZ and global trade compliance product strategy. The Dimensional Research survey referenced in this article was commissioned by QAD, powered by AWS, and surveyed 301 VP and C-level executives in trade compliance, supply chain, and operations.