AI-Powered Document Automation: How IDEXX Uses cbs AID to Improve Accuracy

Many organizations still rely on manual entry across finance, supply chain, and compliance workflows, even as document volumes and complexity increase. However, some are beginning to turn to AI-powered document automation, which converts complex, human-readable documents into structured data that feeds directly into SAP processes.

At SAPinsider Las Vegas 2026, Bastian Schiele, Senior Manager – Technology at cbs Corporate Business Solutions, and Prakash Sethia, Enterprise Architect at IDEXX Laboratories, presented a session on how AI-driven automation can transform documents into SAP-ready data across SAP ECC, SAP S/4HANA, and SAP Global Trade Services (GTS).

The session focused on practical adoption, including how organizations can start with high-impact use cases and scale based on measurable results.

cbs AID Converts Documents into SAP-Ready Data

The solution presented by cbs is called cbs AID (Advanced Integration of Documents).

It is designed to move documents directly into SAP processes, replacing the manual step where users read documents and enter data into the system. Instead, documents are ingested, processed, and converted into structured data that can be posted into SAP.

The workflow follows a defined sequence.

Documents enter through channels such as email, API, or user upload, then pass through preprocessing steps to make content readable. From there, the system extracts relevant data and maps it into a target schema aligned with SAP objects.

“We are not just extracting the information… we integrate that seamlessly into the end-to-end process,” said Schiele, describing how the output feeds directly into SAP transactions such as orders, deliveries, or financial records.

The system is built to handle document variability. Contracts, invoices, and certificates often include tables, handwritten inputs, and inconsistent layouts. The solution applies specialized processing to interpret these structures and identify relevant data points.

Validation is a core component. “No AI can be 100% correct,” Schiele said, emphasizing that the system uses confidence thresholds, backend checks, and human review to ensure that uncertain data is flagged before it enters SAP.

This approach positions document automation as part of the execution layer in SAP, where data is interpreted, validated, and integrated before it becomes part of a business process.

IDEXX Applies AI Document Automation in SAP Finance

The IDEXX implementation shows how this approach applies in a finance context.

IDEXX, which is a global leader in pet healthcare, processes large volumes of customer contracts that drive revenue recognition in SAP, requiring data to be extracted and entered into SAP Revenue Accounting and Reporting (RAR).

Their challenge is both scale and complexity. “We get hundreds of customers’ contracts in different layouts… different languages… different formats,” said Sethia, describing a process where finance teams manually extracted data points and entered them into SAP.

Manual processing introduced measurable risk. Finance teams manually extracted contract data and entered it into SAP, creating opportunities for errors in calculation and data entry that could affect downstream processing.

The complexity of the documents made automation difficult. “Some of the data points… could be on page 1… could be on page 3,” Sethia said, highlighting the need to interpret context across sections rather than extract isolated values.

IDEXX had previously attempted to solve this problem. Two earlier solutions were abandoned because they did not meet requirements for accuracy or scale, highlighting the need for a more reliable way to interpret contract data.

The company ultimately implemented cbs AID to address these challenges. The system extracts contract data, validates it against SAP backend systems, and maps it into the structures required for revenue processing.

Results were measured early in the process. In the feasibility phase, the system achieved approximately 98% accuracy in extracting and validating contract data across a set of tests.

Implementation was also fast. “We implemented the whole product within a month,” Sethia said, with technical go-live in November and business use beginning in December.

The solution created controlled workflow, where contract data is interpreted, validated, and entered into SAP with reduced error and less reliance on manual processing

Document Automation Expands Within SAP Architecture

The session positioned document automation as an extension of SAP architecture. Building on the IDEXX example, the approach places document processing within existing SAP workflows, where data is interpreted and validated before it enters the system.

The solution runs on SAP Business Technology Platform as an extension layer, integrating with existing systems while avoiding changes to the core ERP environment. This allows organizations to introduce automation without waiting for system transformation.

“You do not need to be on S4/Hana…you can begin this on ECC,” said Connor Payne, Account Executive at cbs Corporate Business Solutions, describing how the approach can be applied across existing SAP landscapes.

The model extends beyond a single use case.

“We have seen a bunch of different use cases… customer orders, order confirmations… invoices… certificates,” said Schiele, pointing to scenarios where organizations extract data from unstructured documents and move it into SAP processes.

The volume of these workflows is a common constraint.

“We have tons of incoming documents… where end users are manually transferring data from PDFs or paper into SAP,” Schiele said, highlighting how document-driven processes continue to shape how data enters SAP systems.

These scenarios share the same structural challenge.

Documents arrive in different formats, often with mixed content and embedded logic, requiring systems to identify relevant sections and interpret them correctly. cbs AID addresses this by breaking documents into components and applying targeted processing.

Document automation becomes a repeatable pattern in SAP environments, where the focus shifts from capturing document data to controlling how that data is structured, validated, and integrated into business processes.

What This Means for SAPinsiders

• Document processing defines data quality. Errors introduced at the document stage propagate into finance, supply chain, and compliance workflows. That makes document automation less about efficiency and more about establishing reliable inputs for downstream SAP processes.
• High accuracy requires system-level validation. The IDEXX case shows that strong results are not driven by extraction alone. Accuracy improves when extracted data is validated against SAP systems and business rules before it is accepted into financial processes.
• AI adoption starts at the process layer. Organizations are introducing AI within specific workflows rather than across entire systems. Document automation shows how targeted use cases can deliver measurable impact without requiring broader architectural change.