About Blue Bridge

Blue Bridge is one of the leading IT services and solutions providers in Baltics, part of Blue Bridge Group. Their clients include both large private and public organisations. Blue Bridge specializes in complex IT infrastructure solutions with minimal downtime and excellent on-demand support, delivered 24/7.

The Challenge

Since one of Blue Bridge key objectives is to minimize any downtime risks and provide extraordinary IT infrastructure support, reducing operational costs and improving response and solution time are also very important factors.

Thus, Blue Bridge has chosen us to introduce high-accuracy and high-performance ITSM ticket assignment. Our specialization in ML and NLP helped us complete this challenging objective.

The primary objective of the solution was to reduce operational costs.

Execution

In collaboration with Blue Bridge staff, our first goal was to complete a statistical analysis of recurring key phrases which are the most relevant in each team's services. As a joint effort, both parties have conducted a throughout analysis of the existing ITSM tickets and their content. Specialization terms and attributes for every possible combination of teams and services have also been identified and implemented.

Apart from continuous improvement of the language models powered by Machine learning, our task was also to address complex business rules, structurize all the available data inside & outside of the ITSM platform, and, consequently, implement models which can adapt to ever-changing organizational structure and various conditions controlled by ITSM administrators.

Results

Here are the highlights of the solution:

• Autonomous ticket identification and assignment allows the customer service to operate 24/7 and reduces operational costs since a human operator is not longer required.
• High average ticket assignment accuracy. Measured and continuously improved daily.
• Continuous learning. Since the solution gets slammed by tons of new inbound emails every day, it needs to have a sophisticated mechanism for unsupervised learning.
• Convenient feedback loop. An example scenario: 200 emails went in during the day, 20 of them were assigned incorrectly. Since our goal was to improve the language processing system in measurable iterations, quality feedback from ITSM admins themselves is very important. This is why we've implemented a convenient feedback loop that's fully integrated within the ITSM platform and communicates with the advanced dashboard where all the text processing and analytics stay.
• Regression prevention mechanism. As the accuracy of the platform's results is measured with each iteration, we are able to automatically identify and prevent any downgrade scenarios and schedule the language models for extra training.
• Flexible configuration of the models training - weights, ratios, prioritized keywords and other parameters management. All within a single dashboard with real-time notifications and sophisticated analytics modules.
• Full integration with the customer's ITSM platform - Ivanti.

Summary

Our autonomous ITSM ticket assignment solution has proven to be very beneficial for increasing the efficiency of Blue Bridge customer support - increasing both response & resolution time, plus, automating different ITSM workflows at day and night with high accuracy service.

More features planned. Stay tuned.