About this Demo: This page demonstrates an advanced Self Learning AI Business Central Agentโ€” that interprets your natural language questions, determines intent, and orchestrates Business Central API calls to deliver precise, schema-aligned answers. This agent is designed to empower business users and analysts to access customer, invoice, and sales order information without technical expertise.
It improves itself by analysing the queries it can't answer and then creates that capability subject to rules. Note: This is a prototype agent, actively evolving, and not a production system. It demonstrates robust multilingual support, error handling, and schema awareness, but may have limitations or unexpected behavior as development continues. Your feedback is welcome as we refine and expand its agent capabilities!
Site Logo
โ† Back to PTEK Home

AI Business Central Agent

Instantly retrieve precise, schema-aligned answers from your Business Central data using an agent.
Ask questions about customers, invoices, and sales orders and get accurate, context-specific answers powered by an AI agent that orchestrates Business Central API calls. This demo shows how an agent can surface your business data for everyone. [Limited functionality has been built for the demo. The architecture is in place which provides a solid base for future enhancement.]

Ask the Business Central Agent:

Maximum 100 characters. Please ask specific questions about your Business Central data.
Multilingual Support

Sample Business Central Data

Here we are using sample customer, invoice, and sales order data for a fictional company (CRONUS UK Ltd.). This content is for demonstration only. Review and ask business-related questions!

๐Ÿงพ View Sample BC Data

How the Agent Works

The dataset is live and sourced directly from Business Central - no snapshots, no time-delays. Your question is processed by an AI agent that uses Azure OpenAI's GPT models to interpret your intent, maps it to the correct Business Central API action, and synthesizes a user-friendly answer. The agent accesses live system data, always returning a clear, schema-aligned response.