Head of Intelligent Automation, Vivicta
The challenge
A large enterprise operating a high-volume supply chain received more than 330 purchase order confirmations daily from suppliers. The confirmations arrived via email in multiple formats, including PDFs and images, requiring significant manual effort to review and process.
Traditional automation approaches proved difficult because suppliers used highly variable document structures. Onboarding new suppliers often required custom development work, increasing costs and slowing scalability.
The solution
Vivicta designed and implemented an intelligent automation framework that combined Generative AI and Robotic Process Automation. Azure OpenAI was used to extract key order information from supplier documents, regardless of layout or format, while UiPath orchestrated the end-to-end workflow.
The solution automatically monitored incoming emails, processed attachments, validated extracted data, and updated SAP. Business users received automated updates, while exceptions were routed for human review to ensure process integrity and governance.
Intelligent automation servicesAbout the customer
The customer is a global organization with a complex supply chain and a large network of suppliers. Efficient and accurate order processing is critical to maintaining operational continuity and ensuring timely deliveries.
To support growth and increasing transaction volumes, the organization sought a scalable approach for handling supplier communications while reducing manual workload and operational risk.
Key benefits
98% extraction accuracy
AI-powered document processing achieved 98% data extraction accuracy, enabling reliable automation of previously manual tasks.
Faster order processing
Automated handling of supplier confirmations reduced manual effort, minimized bottlenecks, and accelerated order processing workflows.
Scalable supplier onboarding
The GenAI-based approach removed dependence on document templates, enabling faster onboarding of new suppliers with minimal IT effort.
