At BTC AG, creating process documentation is an integral part of the standardized automation process. Why? Not only does it help the specialist department and the implementation team to reach a common understanding of all the important requirements for the automation to be created, it also serves as a reference work for the future, for example, if supposed errors occur and the technical context is no longer immediately apparent. Creating high-quality process documentation requires effort, but this can be reduced by using an AI-supported solution, especially in quality assurance prior to acceptance by the client.
The process of AI-based verification
First, the documentation created undergoes an internal AI-supported review process. This ensures that the documentation meets internal quality requirements. The system analyzes the documentation using language models and machine learning to check for completeness and compliance.
A UiPath Agent AI is used to orchestrate the entire review process. This agent is trained to understand the specific requirements for process documentation and to use the necessary tools and models for review.
The AI-based review covers the following aspects:
- Verification of key data: A check is performed to ensure that all essential process details, such as key figures, responsibilities, and other master data required in the specification, are included in the documentation.
- Handling exceptions (business rule exceptions): The agent searches for sections that describe how exceptions are handled. It checks whether all potential errors and exceptions that may occur during process execution are described in the document and handled appropriately.
- Process representation analysis: The agent checks whether all mandatory diagrams and process representations are available in the specified format.
- System description: A check is performed to ensure that all systems and applications affected by the automated process have been described completely and correctly.
Deviations are reported back to the business analyst by the UiPath Agent with recommendations for improvement. Once the UiPath Agent AI has completed its internal review, the documentation is submitted to the client for review and approval. The upstream AI review ensures that the client receives high-quality, complete documentation, which significantly speeds up the approval process and reduces the likelihood of errors.
Would you like to learn more about this use case? Then let’s talk about it together!
You can find insights into further cases in the public sector on the BTC AG website.
