Overview

The business challenge

• Sales order entry is largely manual, with orders received via emails and phone calls
• CSRs manually validate and extract order details from unstructured inputs (emails, attachments, calls)
• Manual mapping of customer details, product codes, and quantities increases dependency on human effort
• High reliance on CSR judgment to ensure order completeness and accuracy
• Lack of automation creates inefficiencies in order processing and ERP order creation

Impact of the problem

PoC objective

• Support CSRs with AI-assisted transactional processing of incoming sales orders
• Improve accuracy of captured sales order information
• Reduce manual effort in order entry process
• Enable efficient processing of orders from email and phone channels

The Sonata solution – what we did

Sonata deployed an AI-powered intelligent agent integrated with Microsoft D365 F&O to automate the sales order entry lifecycle for US Additives. The agent operates across two primary channels – email ingestion and phone-call conversational assistance – and orchestrates business rule validation, CSR approval workflows, and automated ERP posting.

Phase 1 – Email order channel

• AI agent monitors the shared mailbox and identifies incoming emails containing new sales order requests, even when subject lines are vague or non-standard.
• Extracts structured order data from email body and attachments (PDF, Word, plain text).
• Automatically sends a branded acknowledgement email to the customer confirming receipt, with the original email attached.
• Appends all identified orders to a shared staging list in D365, updated with status (success /failed/incomplete).
• Applies business rule validation: customer identification (by email, phone, or customer ID), product code mapping against catalog, and quantity unit normalization to kilograms.
• Notifies CSR for validation and approval via the shared list view before ERP posting.
• On CSR approval, creates the confirmed sales order in D365 F&O and removes the entry from the staging list.

Phase 2 – Phone call conversational channel

• CSR activates the AI conversational assistant during an incoming customer call.
• Agent provides real-time look-up of customer purchase history and account data from D365 to support the CSR during the conversation.
• CSR inputs order details via the conversational interface; the AI agent checks order completeness and prompts the CSR for missing information
• Same validation, staging, approval, and ERP posting workflow as the email channel applies.

Pilot dashboard

• A one-page summary dashboard (manually created in the pilot phase, if required) provides visibility into:orders received per day/week/month
• Successfully posted to ERP
• With incomplete customer-provided data
• Failed during validation or ERP posting

Microsoft technologies used

Microsoft Dynamics 365 Finance & Operations (D365 F&O)

Microsoft Azure AI Services

Microsoft Copilot Studio/Power Platform

Microsoft Azure Logic Apps / Power Automate

Current scope vs future enhancements

Why Sonata

Results and benefits