Client:
MyCom Solutions, s.r.o. – a small company specialising in computer security and server operation
Challenge:
MyCom is facing a growing volume of data from its customers’ security logs and is looking for ways to streamline the work of analysts in detecting anomalies. The goal was to verify the possibility of using machine learning and artificial intelligence to recognise security threats beyond predefined rules and to support the preparation of dynamic filters.
Solution:
As part of EDIH Northeast Bohemia, we implemented a proof-of-concept:
- connection to security log data
- analysis of historical events and basic anomaly detection using AI
- design and application of filters for the analytical tool
- validation that the model is capable of automatically identifying suspicious events
The customer then initiated the continuation of the project with the aim of developing a chatbot that would allow security analysts to communicate with the tool in natural language – e.g. by entering a query as to whether there had been any suspicious logins in a given time window.
Result:
After successfully verifying the basic concept, the company, in collaboration with the EDIH project team and a university partner, took on another challenge – preparing a proof-of-concept for a comprehensive solution. Development of the solution using AI and chatbots continues with the aim of full integration into practice.
Reference:
‘I appreciate the cooperation with the experts from the EDIH NEB team, thanks to which we were able to test the technologies for data processing and security threat pattern detection and verify that it makes sense to continue in this area and invest time and money in preparing a prototype of a more comprehensive solution.’
Lukáš Vejman, Managing Director
