How to Choose Customs Broker Software

The customs brokerage software market has expanded rapidly. Where brokerages once had a handful of ABI systems to choose from, there are now dozens of platforms promising AI-powered automation, faster entry processing, and compliance intelligence. The challenge is no longer finding software - it is figuring out which solution actually fits your operation.

This guide covers the features that matter most, the tradeoffs between implementation approaches, questions to ask vendors, and warning signs that should give you pause. Whether you are evaluating your first automation platform or considering switching from one you have already tried, the framework here will help you make a more informed decision.

Why Choosing the Right Software Matters Now

Three trends are converging to make customs broker software selection more consequential than it has been at any point in the past decade.

Tariff complexity has accelerated. The layering of Section 232, Section 301, and IEEPA tariffs on top of standard duty rates has created calculation challenges that are difficult to manage manually. A single entry can involve multiple Chapter 99 headings stacking on top of each other, and getting it wrong exposes your importer and your brokerage to significant financial liability. You can model stacking scenarios using the Cervo tariff simulator, but handling them systematically across thousands of entries requires software that understands the logic.

Volume is growing faster than headcount. Finding experienced entry writers and compliance staff is harder than ever. The brokerages that are scaling successfully are automating the repetitive portions of entry preparation - document extraction, routine classifications, standard compliance checks - so their licensed brokers can focus on exceptions and client relationships.

Compliance requirements keep expanding. PGA message sets have grown more complex, AD/CVD cases cover more products, and denied party screening lists are updated more frequently. The compliance surface area that brokers need to cover on every entry has widened, and doing it manually creates both speed and accuracy problems.

The right software addresses all three of these pressure points. The wrong software adds a new set of problems on top of the ones you already have.

Key Features to Evaluate

Not every feature matters equally for every brokerage. But these are the capabilities that consistently determine whether a customs automation platform delivers real value or becomes another tool your team works around rather than with.

Document Extraction and Data Processing

This is the foundation. If the software cannot reliably extract data from commercial invoices, packing lists, bills of lading, and certificates of origin, nothing downstream works well. When evaluating document processing capabilities, focus on these specifics:

• Format flexibility: Can it handle PDFs, scanned images, Excel files, and email body text? Your importers do not send documents in a consistent format, and the software needs to handle that reality.
• Multi-page and multi-invoice handling: Many commercial invoice PDFs contain multiple invoices or combine invoices with packing lists. The system should parse these correctly without manual splitting.
• Field-level accuracy: Ask for accuracy rates broken down by field - shipper, consignee, HTS, unit price, quantity, country of origin. Overall accuracy percentages can mask poor performance on critical fields.
• Handling of non-standard layouts: The real test is not whether the system works on clean, well-formatted invoices. It is whether it works on the messy ones - handwritten notes, foreign language descriptions, merged cells, inconsistent units of measure.

HTS Classification Automation

Automated HTS classification is where most customs automation platforms differentiate themselves, and it is also where accuracy claims require the most scrutiny. Key evaluation criteria include:

• Commodity coverage: Some platforms perform well on consumer goods but struggle with industrial components, chemicals, or textiles. Test with your actual commodity mix, not a generic demo dataset.
• Confidence scoring: The system should provide a confidence level with each classification and route low-confidence items to a human reviewer. A platform that classifies everything with equal confidence is not being honest about its limitations.
• Learning from corrections: When your team overrides a classification, does the system learn from that correction? A  or similar mechanism that stores validated classifications for recurring items is essential for improving accuracy over time.
• Chapter 99 and special programs: Classification is not just about the base HTS code. The system needs to determine applicable Section 232, Section 301, IEEPA headings, and other Chapter 99 provisions. This is where many platforms fall short.

Compliance Screening

Compliance screening covers several distinct checks that need to happen on every entry. Evaluate the platform's capabilities across each of these areas:

•  Does the system identify when goods require Partner Government Agency involvement - FDA, USDA, EPA, CPSC, FCC, and others? Does it generate the correct PGA message set codes, or does it just flag that a PGA might be involved?
• AD/CVD screening: Can the system cross-reference HTS codes and countries of origin against active  orders? This is a critical compliance check that carries serious penalties when missed.
• Denied party screening: The system should screen against all relevant lists - the Consolidated Screening List, SDN list, Entity List, and others - and do so against all parties on the entry, not just the shipper.
• Free trade agreement eligibility: Can the platform identify when goods may qualify for preferential duty treatment under USMCA, CAFTA-DR, or other trade agreements?

Rules Engine and Per-Importer Customization

Every brokerage has importers with specific requirements - preferred HTS codes for recurring products, specific document handling instructions, custom compliance checklists, or unique data formatting needs. The software needs to accommodate this without requiring custom development for each importer.

Look for a rules engine or configuration layer - sometimes called an agent studio or workflow builder - that lets you define importer-specific logic. This might include routing rules for document review, override hierarchies for HTS classification, or automated handling of specific commodity types. The more you can configure without writing code or submitting support tickets, the more value you will get from the platform.

ABI Integration

This is often the most underappreciated evaluation criterion. A customs automation platform is only as useful as its ability to get data into your filing system. Key questions include:

• Which ABI systems are supported? CargoWise, OneView, NetCHB, and DesCartes are the most common. Confirm that the platform has a working, production-tested integration with yours - not just an API that theoretically could connect.
• What data flows through the integration? Does it push complete entry data including line items, values, HTS codes, and PGA information? Or does it only push header-level data, requiring your team to manually enter line details?
• Bidirectional or one-way? Can the automation platform read data back from your ABI system to confirm filing status, pull entry summaries, or sync updates? Or is data flow one-directional only?
• How is the integration maintained? ABI systems update their APIs and data formats periodically. Ask who is responsible for maintaining compatibility when those changes happen.

Reporting and Analytics

Beyond processing individual entries, the software should give you visibility into your operation - entries processed, automation rates by importer, accuracy tracking showing which fields required human correction, compliance dashboards summarizing PGA filings and AD/CVD flags, and duty analytics by tariff program and country of origin. Good reporting helps you demonstrate value to importers and identify opportunities like drawback recovery.

Build vs Buy vs Layer

One of the most important strategic decisions in customs automation is not which vendor to choose, but which approach to take. There are three fundamentally different paths, and each carries distinct tradeoffs.

Option 1: Build In-House

Some larger brokerages consider building automation tools internally. The appeal is full control and no recurring software costs. The reality is that building customs-specific AI requires deep expertise in both machine learning and trade compliance. Most in-house attempts produce tools that work for a narrow set of document formats, then plateau. Maintaining them as tariff programs change becomes a permanent engineering commitment. For most brokerages, this is not the right path.

Option 2: Buy a Full Platform (ABI Replacement)

Some vendors offer end-to-end platforms that combine automation with ABI filing capabilities, aiming to replace your existing system entirely. This simplifies your tech stack, but the risks are significant: migrating bonds, importer configurations, and historical data takes 6 to 12 months minimum. Your ABI system is the backbone of your operation, and replacing it while continuing to file entries on deadline introduces risk that most brokerages underestimate. And once you have migrated everything, switching again becomes even more expensive - limiting your leverage if the platform underdelivers.

Option 3: Layer Automation on Top of Your Existing System

The layered approach keeps your ABI system in place and adds automation capabilities on top. The automation platform handles document extraction, classification, compliance screening, and data validation, then pushes prepared entry data into your ABI for filing. This approach offers several advantages:

• Lower risk: You are not replacing your filing system. If the automation platform underperforms, your team can fall back to their existing workflow.
• Faster implementation: Without data migration or system replacement, you can be live in weeks rather than months.
• Incremental adoption: Start with one importer or one entry type, validate the results, and expand from there.
• Flexibility: If you later decide to switch ABI systems, your automation layer moves with you.

For the majority of brokerages, the layered approach is the most practical path to automation. It lets you capture the value of AI-powered processing without taking on the risk of a full infrastructure change.

Questions to Ask Vendors

When you are evaluating customs automation platforms, these questions will help you cut through marketing language and get to the substance of what the platform actually does.

On Accuracy

• What is your document data extraction accuracy rate, broken down by field type?
• What is your HTS classification accuracy rate for my specific commodity mix?
• Can you run a blind test on 50 to 100 of my actual commercial invoices during the evaluation?

On Implementation

• How long from contract signature to processing live entries?
• Do I need to change any of my existing processes or systems?
• Can I start with a pilot on a subset of importers before rolling out broadly?

On Integration

• Do you have a production-tested integration with my specific ABI system?
• What data fields flow through the integration, and are there any that require manual entry?
• Can I see a reference customer who uses the same ABI system I do?

On Pricing

• Is pricing per entry, per line, per user, or flat rate?
• What happens to pricing as my volume grows?
• What is the contract term, and is there an exit clause if performance does not meet expectations?

Red Flags to Watch For

In any software evaluation, there are signals that should make you proceed with caution. In the customs automation space specifically, watch for these warning signs.

Requiring a Full Rip-and-Replace

If a vendor insists that you need to replace your ABI system to use their automation capabilities, ask why. There are legitimate architectural reasons why some platforms work this way, but it significantly increases your implementation risk, timeline, and switching costs. Unless you are already planning to change ABI systems for other reasons, a rip-and-replace requirement should give you pause.

No Pilot or Trial Option

Any vendor that requires a long-term contract commitment before you can test the platform with your actual documents is asking you to take on risk that they should be willing to share. A confident vendor will let you run a pilot, process real entries, and validate accuracy before you sign a multi-year deal. If a vendor will not offer a trial period, ask yourself what they are worried you will discover.

Vague Accuracy Claims

Phrases like "99% accurate" or "industry-leading accuracy" without context are meaningless. Accuracy depends on what is being measured (data extraction vs classification vs end-to-end entry accuracy), on what types of documents and commodities, and against what baseline. Credible vendors provide specific, verifiable metrics and are transparent about where their platform performs well and where it struggles.

No Customs Domain Expertise on the Team

If the vendor's team does not include people with direct customs brokerage experience - licensed brokers, former entry writers, compliance specialists - the platform is likely to struggle with the edge cases and regulatory nuances that define real-world customs work. General-purpose AI is not the same as customs-specific AI.

Unclear Handling of Regulatory Updates

Tariff programs, HTS codes, and PGA requirements change regularly. Ask how quickly the platform incorporates new Chapter 99 headings, updated AD/CVD orders, or revised PGA message set requirements. A platform that lags on regulatory updates creates compliance exposure for your brokerage and your importers.

How Cervo Fits

Cervo AI was built specifically for customs brokerages, and its architecture reflects the layered approach described above. Rather than asking you to replace your ABI system, Cervo connects to the systems you already use - including CargoWise, OneView, and NetCHB - and handles the automation layer on top.

Here is how Cervo maps to the evaluation criteria in this guide:

• Document processing:  extracts data from commercial invoices, packing lists, and related documents across formats including PDFs, scanned images, and spreadsheets. Accuracy is measured and reported at the field level.
• HTS classification:  with confidence scoring, human review routing for low-confidence items, and a  that stores validated classifications for recurring products.
• Compliance screening: Integrated , AD/CVD cross-referencing, and denied party checks on every entry.
• Rules engine:  lets you configure importer-specific workflows, classification overrides, and compliance rules without writing code.
• ABI integration: Production-tested integrations with CargoWise, OneView, and NetCHB. Entry data including line items, values, HTS codes, and PGA information flows directly into your filing system.
• Entry automation: End-to-end automation for , , , and .

Cervo offers a pilot program that lets you process your actual entries before making a commitment. You can also explore tariff stacking scenarios using the free tariff simulator to understand how Section 232, 301, and IEEPA duties interact for your commodity mix.

Frequently Asked Questions

Should I replace my ABI system or layer automation on top of it?

For most brokerages, layering automation on top of your existing ABI system is the lower-risk approach. Replacing your ABI means migrating bonds, importer records, billing configurations, and operational workflows all at once. A layered solution lets you automate document processing, HTS classification, and compliance screening while keeping your existing system of record intact. This approach also lets you pilot with a subset of entries before committing fully.

What accuracy rate should I expect from AI-powered customs software?

Credible platforms should be able to demonstrate 95% or higher accuracy on document data extraction and provide specific metrics broken down by document type. For HTS classification, accuracy depends heavily on the commodity mix - a platform might achieve 98% on apparel but 85% on industrial machinery. Ask vendors to run your actual documents through their system during evaluation rather than relying on aggregate accuracy claims.

How long does it take to implement customs automation software?

Implementation timelines vary significantly depending on the approach. Full ABI replacements typically take 6 to 12 months including data migration, user training, and bond transfers. Layered automation solutions that connect to your existing ABI system can be operational within 2 to 4 weeks, since they do not require replacing your core infrastructure. The key factor is whether the solution requires changing your filing system or simply feeds data into it.

What is the difference between customs automation software and an ABI system?

An ABI system like CargoWise, OneView, or NetCHB is your system of record for filing entries with CBP. Customs automation software handles the upstream work - extracting data from documents, classifying goods, screening for compliance, and preparing entry data. The automation layer feeds validated data into your ABI system so your team spends less time on manual entry. Some platforms try to do both, but many brokerages find it more practical to keep their ABI system and add automation on top.

See How Cervo Fits Your Brokerage

Every brokerage has a different mix of importers, commodities, and system requirements. The best way to evaluate Cervo is to run your actual documents through the platform. Request a demo and we will walk through your specific workflow - including document processing, HTS classification, compliance screening, and integration with your ABI system.

Get a Demo