Summary
Itacon aimed to digitize and scale its admission process but faced challenges due to technical debt and lack of flexibility in their current solution. To address these issues, we developed an AI expert system for automating the evaluation of security audit questionnaires. Using tools like Python, OpenAI, and Scikit-Learn, we enabled the system to provide clear and accurate responses, significantly improving transparency and efficiency. The AI system reduced checklist completion time to 20 minutes with 64% accuracy, streamlining the security audit process.
Challenge
Motivation to change
RESIDE wanted to expand, scale-up, and roll out a complete digitalization of the admission process in the US. However technical debt and shortcomings in the configurability and flexibility of the solution didn’t allow for that. Reside was also struggling with the transparency of the software development process and predictability of the delivery.
Solution
Change delivery
The goal of the project is the development of AI expert system responsible for evaluation of security audit questionnaire based on supplied client documents.
Tools & means
- Python
- OpenAI
- LangChain
- Mlflow
- ScikitLearn
- Ragas
Outcomes
Change outcomes
Our assistance was very helpful in completing their security audit questionnaire. The AI provided clear, accurate answers that streamlined the entire process. It handled complex security questions effectively, saving them time and ensuring their responses were thorough and precise.
Used services
-
64%accuracy
-
20 minutescompletion of 1 checklist