01 JULY 2026
Welcome to the Genpact Deductions Recovery Release Notes.
What's in the Release Notes?
- About Genpact Deductions Recovery
- Main features
- Enhancements
- Known issues
- Limitations
About Genpact Deductions Recovery
Genpact Deductions Recovery is a GenAI and ML supported standalone independent solution that provides end-to-end capability to process deductions. The Genpact Deductions Recovery solution collects data from the ERP system, carrier portal and customer portal, validates the data, and provides deduction management to process the data for valid or invalid deductions.
Efficient deduction management with Genpact Deductions Recovery enhances cash flow, reduces revenue leaks, strengthens partnerships, and helps in business sustainable growth.
Main features
- Out-of-the-box support for North America CPG retailers: In addition to Amazon (released in the beta version), the solution now supports six of the top North America CPG retailers, ADUSA, Walmart, Costco, Home Depot, Dollar General, and Target.
For the seven retailers, the solution offers the following:- No document onboarding is required for the extraction service.
- The default validation rules in the system already know how to validate each retailer’s data with no configuration or modification, thus giving operations teams a true plug-and-play experience.
- Optimized extraction with Extraction as a Service (EaaS) solution: The Genpact Deductions Recovery integrates with Extraction as a Service (EaaS) to offer accurate extraction of header and line-level data from the documents.
EaaS is an AI-powered intelligent document processing platform that exposes document classification, splitting, and data extraction capabilities as ready-to-consume APIs. Using EaaS for extraction adds the following values to the system:
- Accelerated growth and onbaording: The solution offers a configuration-driven extraction approach that expedites the onboarding of new retailers, documents and deduction types.
- Improved accuracy: The solution combines large language models (LLMs), Azure AI services (Content Understanding and Document Intelligence), and open-source OCR to convert unstructured and semi-structured documents into reliable, structured data.
- Agentic and automation capabilities: The solution supports agentic investigation, validation and resolution, thus moving deductions towards low-touch orchestrated workflows.
- Reduced complexity: The solution separates platform extraction issues from system domain logic.
- De-risked SaaS scale: The solution provides a SaaS extraction layer that scales predictability without client-specific reimplementation or operational risk.
- Support Packing List data: The system now supports the Packing list documents data for deductions.
Packing list support is optional and is enabled per delivery through configuration. When enabled:- A background process pushes the packing list data flat file into the database.
- The packing list copy is added to the case as an attachment for reference.
- Packing list data can be added to the validation rules per business request. It is not included in the default validation rules.
- Extended Exception Handling: The Genpact Deductions Recovery system supports manual exception handling to allow an analyst to manually handle case exceptions in the following scenarios:
- When the system is unable to complete the processing cycle or line-items validation.
- When low‑quality files result in incomplete or inaccurate data extraction.
- When the Proof of Delivery (POD) contains notations that require human interpretation, such as overages, returns, refusals, or similar exceptions.
For details, see this article and this article. - Upload deduction case files in one submission: We have introduced two new forms for users to upload multiple case files in one submission without needing direct blob access. Analysts retrieve all files manually and upload them through one of the two forms:
- Cross-case files: For bulk data files such as the FBL5N report, VBRP invoice data, packing list data, and the Dollar General remittance file.
- Per-case files: In-case files such as the invoice copy, POD, backup, and packing list.
For details, see this article.
Enhancements
- Enhanced Validation: The system offers a four-way matching process using YAML rules to validate the deduction case line-items as valid or invalid or partially valid. We have enhanced the validation rules with the following:
- Additional fields from the source files that can be used in the validation rules per business requirements.
- Data normalization to correctly convert and compare units of measure (UOM).
- Logic to handle cases that require human intervention.
- Option to use the packing list data in the validation rules per business request. However, packing list data is not part of the default rules.
- Database and performance improvements: As part of performance improvement, the system now supports additional SAP fields.
Known issue
- The system currently doesn't support TIFF and TIF image files as input documents.
Limitations
- EaaS can't support, in a single submission, a file larger than 18 MB.
- All the structured files (Report, Backup, VBRP) must not contain any blank or empty column names.
- All structured files must be from XSLX or CSV file types to be processed by ADF.
