Through workshops with the customer, we discussed how artificial intelligence (AI) could be used to help customers as a first step in solving the problem. The focus was on how to quickly get started and learn along the way in the process. During the workshop, three areas were selected for further work in a Proof-of-Concept (PoC) phase. The three areas selected were:

‍

1. Match invoices with payments in the database

2. Receive unpaid invoices and send emails to customers

3. Develop a database agent to retrieve information for non-technical users

‍

Over the course of one month, we developed a Proof-of-Concept for each of these areas. Based on the experience gained, we decided to make the solutions for points 1 and 2 production-ready for the enterprise. A matching algorithm was developed and set to run once per 24 hours, while at the same time a KI agent was developed to retrieve invoices and email them to customers.

This laid the foundation for further development of the agent so that it could also find relevant information in the database, retrieve correct files, send emails and generate reports.

Going forward, the goal is to continually add new tasks for the agent to perform, so that it ultimately functions as a technically competent employee who is available to all employees 24/7.

‍

A KI-powered platform with a common user interface that streamlines information and file management, as well as communication with customers. The user can easily interact with an AI agent who acts as a digital colleague, performs tasks, simplifies workflow and increases productivity. The agent retrieves files from various folders, extracts relevant information from databases, and generates email drafts with attachments, so that the user can quickly send the necessary information to the appropriate recipient.