Two women in storehouse with packages on conveyor belt

Case study

How to automate invoice validation and modernize tax with GenAI

One logistics leader used EY.ai for tax, built with IBM watsonx®, to reduce manual effort and focus on strategic priorities.

 

Many logistics providers are rapidly embracing digital transformation and emerging technologies to modernize operations and meet rising customer expectations in a highly dynamic global supply chain environment.

 

For one such organization — a global industry leader — maintaining their competitive edge means investing billions of dollars in capital programs each year, including building distribution centers, expanding transportation fleets, and implementing advanced technologies to modernize operations. As a result, the tax professionals tasked with reconciling invoices from vendors involved in these initiatives — as well as those from day-to-day-operations — continuously wrestle with overwhelming administrative burden because:

 

  • Invoices vary in format depending on the supplier, jurisdiction, and country-specific tax requirements — adding significant complexity to what should be a repeatable process.
  • Regulatory changes often come with new filing requirements, reporting formats, or documentation standards.
  • Tax regulations shift frequently, and discrepancies in invoice data can go undetected due to compressed month-end timelines.

 

With highly skilled resources spending entire weeks manually reviewing invoices, they were unable to focus on more strategic endeavors. The process also introduced compliance risks and delays in financial reporting. “I knew that Generative AI could help drive increased effectiveness and efficiencies within our tax function,” said the logistics provider's head of tax. “But we lacked the tools and AI expertise internally to move forward.”

 

Plug in intelligence, preserve core systems

 

Working together, EY and IBM developed EY.ai for tax, built with IBM watsonx, which is designed to help transform how tax departments function. It addresses the data-related challenges that can limit workflow automation, supports advisory activities on demand, and helps tax teams prioritize value-added activities over routine processes.

 

The AI engines, automation tools, and advanced analytics within EY.ai for tax, built with IBM watsonx, can be deployed seamlessly on top of existing systems without disrupting them by plugging in via APIs, thereby enhancing the current infrastructure. This means that organizations can avoid the costly, risky, and time-consuming process of replacing legacy systems that weren’t designed with the requirement of AI in mind.

 

EY.ai for tax, built with IBM watsonx, at a glance:

  • Scales tax team productivity and data-driven decision making.
  • Enables teams to focus on higher-value-added activities, such as analyzing and acting on data.
  • Streamlines compliance initiatives and delivers greater business value.

Enterprise-grade GenAI powers smarter tax operations

Optical character recognition (OCR) tools are often used to review invoices with varying levels of first-pass accuracy, and over time machine learning retraining can drive some improvements. With EY.ai for tax, built with IBM watsonx models, first-pass accuracy can increase to 95%.

As an early-stage adopter of EY.ai for tax, built with IBM watsonx, the logistics provider leveraged one of three initial AI-enabled use cases, Detect and Correct with Business Documents, to automate invoice validation against ERP data extracts for tax determinations and filings.

The EY.ai for tax, built with IBM watsonx:

  • Ingests PDFs, images, and electronic data interchange (EDI) feeds directly from existing purchase-to-pay workflows.
  • Uses large-language models running in watsonx.ai on IBM Cloud to classify, extract, and validate data.
  • Maps results to master records through EY accelerator libraries.
  • Routes only true exceptions to an insight dashboard for rapid human review.

The pilot revealed that EY.ai for tax, built with IBM watsonx, was able to review the tens of thousands of vendor invoices with a 95% first-pass accuracy rate. What’s more, the tax team, as well as accounts payable and audit support, were able to redirect their focus toward higher-value, strategic activities without disruptive system overhauls.

Meaningful results with EY.ai for tax, built with IBM watsonx

Before

After

Multi-day manual validation delays monthly close.

Auto-validation of invoices shortens close cycle. 

Inconsistent application of regional tax rules across business units.

Centralized AI logic standardizes tax treatment enterprise wide.

Tax professionals consumed by vendor invoice reviews.

Reclaimed hours redeployed to tax planning and controversy.

Proven win builds confidence for an AI-enabled future

As a result of the transformative benefits realized with the initial use case, the logistics provider is excited to implement the EY.ai for tax, built with IBM watsonx Contract Review Agent, which scans purchase, construction, and fleet-lease agreements at the front end of the process. This additional capability will enable the company to intercept tax exposure earlier, enrich the growing tax data lake, and accelerate the journey toward a fully AI-enabled tax lifecycle.

Take the next step on the AI adoption journey 

In an environment where accuracy, speed, and regulatory alignment are critical, legacy approaches to tax compliance and reporting are a bottleneck to both operational efficiency and business agility. Wherever organizations are on the AI adoption journey, EY.ai for tax, built with IBM watsonx, makes it possible to leverage the power of GenAI — no rip and replace required.