Avalara Automated Tariff Code Classification streamlines the process of assigning tariff codes to products, ensuring compliance with trade regulations and optimizing import/export operations.
Avalara Automated Tariff Code Classification is an AI-driven solution that automates the assignment of Harmonized System (HS) codes to products at 6-digit or country-specific 10-digit levels, streamlining compliance with global trade regulations and enabling accurate customs duty calculations for import/export operations.[1][2][3] Its core value lies in replacing error-prone manual processes with advanced AI, machine learning, and natural language processing that analyze product titles, descriptions, categories, and destinations, drawing from a vast database of over 30 million classifications and continuously updating via World Customs Organization data.[1][3][4]
Key features include adaptive matching technology for rapid catalog processing via CSV, SFTP, or API; confidence scores for predictions; optional expert validation in Pro tiers; and seamless integration with Avalara's broader tax compliance platform, supporting 180+ countries.[1][2][6] Ideal for supply chain providers, e-retailers, global marketplaces, brokers, shippers, and importers handling large catalogs or complex products, it reduces delays, penalties, and costs associated with inaccurate classifications.[2][3][4]
What sets it apart is its pioneering continuous learning loop—where validated codes enhance the AI models, including customer-specific versions—and hybrid human-AI expertise for superior accuracy on nuanced items like ingredients or materials, outperforming traditional methods in speed and scalability.[1][3][5]
Sources
Avalara explicitly states it will NOT use customer data to train AI models unless expressly permitted in writing. The te
No training on user data means no opt-out is necessary.
Low risk across data training, licensing, and sharing dimensions.
Avalara explicitly states it will NOT use customer data to train AI models unless expressly permitted in writing. The terms clearly protect against unauthorized AI training on user data.
“[ToS] 'Avalara will not use Customer Data submitted through the Services to train or improve its AI large language models or those of third-party providers unless expressly permitted in the Agreement, the Data Processing Agreement, or another written agreement with Customer.'”
Customer retains all ownership rights in Customer Data and Customer Output. Avalara's license is strictly limited to providing the service, with no sublicensing or perpetual rights claimed over customer-generated content.
“[ToS] 'Customer retains all ownership rights in Customer Data, Customer Output, and Customer's Confidential Information, including any Intellectual Property rights therein. Nothing in the Agreement transfers or conveys to Avalara any ownership interest in or to the Customer Data, Customer Output, or Customer's Confidential Information.'”
Avalara may use subcontractors and Affiliates to provide services, and shares data with them under confidentiality obligations. While a DPA is referenced and sub-processors are mentioned, the documents provided do not include the full sub-processor list or detailed DPA terms.
“[ToS] 'Avalara may use subcontractors to facilitate its obligations under the Agreement, and Avalara shall be responsible for the acts and omissions of such subcontractors relating to the Agreement as though they were those of Avalara.' Also: 'Avalara may perform the Services itself or through any of its Affiliates.'”
Avalara provides some clarity on data deletion (60-day window post-termination for customer requests, then destruction per retention policies), but retention periods during active service are vague and tied to 'regulatory, legal, or audit purposes' without specific timelines.
“[ToS] 'Upon Customer's request at any time during the Subscription Term for a Service or up to 60 days after termination of the Subscription Term, Avalara shall promptly destroy or overwrite Customer Data for such Service or Customer's Confidential Information, other than Customer Data or Customer Confidential Information or Personal Information contained in automatic computer backups or historical archives or that must be retained to fulfill obligations under the Agreement for regulatory, legal, or audit purposes.'”
Customer owns all Customer Output (results, calculations, reports generated by the service). Avalara claims no ownership or license rights to customer-generated outputs, only to its own technology and content.
“[ToS] 'Customer Output means all results, calculations, reports, data sets, or other similar materials generated by the Service solely from Customer Data through Customer's authorized use of the Service. For clarity, Customer Output does not include the Service, Avalara Technology, Content, or any improvements, enhancements, or derivatives thereof.'”
The documents reference a Data Processing Agreement (DPA) and GDPR/CCPA compliance, and mention independent third-party audits of security programs, but specific certifications (SOC 2, ISO 27001) are not explicitly stated in the provided excerpts.
“[ToS] 'Avalara uses independent third parties to perform audits of our security program ensuring we have comprehensive controls protecting your data.' Also: 'The Parties agree to comply with the Avalara Services Data Processing Agreement, which is incorporated by this reference and is located at https://www.avalara.com/GDPR-DPA (the "DPA").'”
Track and measure your AI tools with Olakai's observability platform.
Explore Olakai