--- title: "Sainapse: Avery Dennison, Email to Service Request, at 300k/Day" description: "Avery Dennison's 1,000-agent desk takes 300,000+ emails a day. Sainapse reads, classifies, and files each into Oracle Service Cloud; ~5% reach a person." canonical_url: "https://www.sainapse.ai/customers/avery-dennison-email-to-service-request" markdown_url: "https://www.sainapse.ai/customers/avery-dennison-email-to-service-request.md" text_url: "https://www.sainapse.ai/customers/avery-dennison-email-to-service-request.txt" archetype: "customer-story-leaf" --- # Sainapse: Avery Dennison, Email to Service Request, at 300k/Day Customer story ## Sainapse files 300,000+ emails per working day into Oracle Service Cloud Avery Dennison ran a 1,000-agent service desk where every inbound email was read, classified, and keyed into an Oracle Service Cloud service request by hand. Sainapse now reads the email, pulls the order details, and creates the service request itself. The ~5% of cases that need a person still go to one. On this page: - [The challenge](https://www.sainapse.ai/customers/avery-dennison-email-to-service-request#customers_avery_dennison_email_to_service_request_03) - [What Sainapse delivered](https://www.sainapse.ai/customers/avery-dennison-email-to-service-request#customers_avery_dennison_email_to_service_request_04) - [Outcomes](https://www.sainapse.ai/customers/avery-dennison-email-to-service-request#customers_avery_dennison_email_to_service_request_05) - [Tools involved](https://www.sainapse.ai/customers/avery-dennison-email-to-service-request#customers_avery_dennison_email_to_service_request_06) - [Related deployment](https://www.sainapse.ai/customers/avery-dennison-email-to-service-request#customers_avery_dennison_email_to_service_request_07) Updated August 2026 Before Sainapse ## The challenge Before Sainapse, each agent worked through 300+ emails a day, across multiple languages. Every message needed a read, then a call on which service request type it was (rush order, order revision, order monitoring, and others), then the order details pulled out of free text and typed into Oracle. Throughput was capped by how fast people could read, and service request quality drifted between agents and shifts. The intake queue scaled with the business. The team could not. The deployment ## What Sainapse delivered Sainapse classifies each email against the service request types the desk actually uses, then pulls the fields Oracle Service Cloud needs out of the free-text body. Where the email leaves context out, Sainapse checks the Sales Order and Production Planning APIs in real time and fills it in. It creates the service request directly in Oracle Service Cloud, the system the desk already runs on, and it handles mail in multiple languages for global operations. Every agent correction goes back into the model, so accuracy improves with use. Outcomes ## What changed in production. ### 300,000+ emails per working day Read, classified, and filed as service requests, against a baseline of 300+ emails per agent per day. ### ~5% of cases Routed to a person, sustained in production. ### 1,000 agents moved off email intake Onto the problems that need them. ### Multi-language coverage Multi-language processing live across global operations. Product ## What this desk is assembled from. Nothing here was built for Avery Dennison alone. Each piece below is a shipped capability with its own page. ### Document Intelligence - [Workflow overview](https://www.sainapse.ai/workflows/document-intelligence): The Document Intelligence workflow this desk runs on. - [Intelligent Field Extraction](https://www.sainapse.ai/features/intelligent-field-extraction): Pulls the fields Oracle Service Cloud needs out of the free-text body. - [Intelligent Ticket Classification](https://www.sainapse.ai/features/intelligent-ticket-classification): Classifies each email against the service request types the desk uses. - [End-to-End System Write](https://www.sainapse.ai/features/end-to-end-system-write): Creates the service request directly in Oracle Service Cloud. - [Human-in-the-Loop Exception Handling](https://www.sainapse.ai/features/human-in-the-loop-exception-handling): The ~5% of cases that need a person still go to one. ### Explore - [All customer stories](https://www.sainapse.ai/customers): Named deployments and what they measured. Related ## Related deployment ### [Avery Dennison: PO to sales order](https://www.sainapse.ai/customers/avery-dennison-po-to-sales-order) Companion story: Sainapse also runs the PO-to-Sales-Order pipeline at Avery Dennison, writing sales orders into Oracle CPQ on the order operations side of the business. ### [Document Intelligence](https://www.sainapse.ai/workflows/document-intelligence) Workflow: The workflow this deployment runs: format-agnostic reading, field extraction, validation, and a direct write into the system of record. ## Point it at your own inbound mailbox. - [Book a demo](https://www.sainapse.ai/book-a-demo) ## Sitemap See the full [Sainapse sitemap](https://www.sainapse.ai/sitemap.md).