Manual Voice Order Trigger
The workflow is initiated manually when a voice-based customer order is ready to process, launching the AI-powered pipeline that converts call transcriptions into structured Dynamics 365 Business Central sales orders.
About this template
AI-powered voice-to-order automation template converting voice calls to Voice to Draft Order in Business Central orders in Dynamics 365 Business Central
Keep exploring
Every repetitive process between your ERP, CRM, and business apps has a template. Find yours, customize it, and go live today.
Customize the flow
Every module in this template is fully customisable. Add triggers, swap actions, or extend the logic using appse ai's no-code visual builder.
No triggers for HTTP.
Inside the flow
appse ai runs this workflow end to end for HTTP — automatically, every single day. Here is exactly what happens when this template runs.
The workflow is initiated manually when a voice-based customer order is ready to process, launching the AI-powered pipeline that converts call transcriptions into structured Dynamics 365 Business Central sales orders.
A GET request is sent to an external transcription service to retrieve the full voice call transcript, capturing all order-related information spoken by the customer during the interaction.
The transcription is split into individual sentences for granular processing, enabling the system to evaluate each line of the conversation separately for relevant order content.
The system filters sentences using order-related keywords such as 'place' and 'order,' isolating only the customer's purchasing instructions and discarding non-order conversational content.
An OpenAI model processes the filtered transcript to extract structured order data—including customer ID, product names, and quantities - converting natural language into machine-readable JSON for Business Central voice order automation.
Extracted product names are matched against a Qdrant vector database to retrieve precise item details—descriptions, prices, and item codes, ensuring accurate product identification even from informal or abbreviated customer speech.
AI matches extracted order items with Qdrant product records, associating product names and quantities with correct item codes and pricing, producing a fully structured and enriched order payload ready for ERP submission.
The enriched order data is sent to Dynamics 365 Business Central to create a new sales order with all customer and product details, enabling hands-free order processing directly from the call transcript.
If key order details such as customer ID are absent from the transcription, the system automatically sends a Microsoft Teams alert to the relevant team for prompt follow-up, ensuring no order is lost due to incomplete call data.
Discover how appse ai has helped businesses automate processes and achieve growth through our users' success stories.