Digital innovation: chatbots, RPA and ML in business service

Automatic creation of incidents in ITSM using RPA

One of the most common problems for customers is the lack of integration between the ITSM system and the monitoring system. This happens for various, including technical, reasons. The monitoring team works mechanically, regularly checking email notifications in anticipation of new alerts from the monitoring system, and manually creating incidents in the ITSM tool. Software integration of systems is difficult, expensive, or not possible at all.

In order to automate tasks and, as a result, improve the efficiency of the monitoring group, the experts from ICL Services offered customers the option to introduce RPA technology in the monitoring system and provide instrumental integration of systems. Robotic Process Automation (RPA) is a business process automation technology based on RPA robot. This program is capable of performing typical and repetitive (routine) operations, which person performs using the user interface.

As a result, the following is possible in real projects:

• Automatically creating incidents in ITSM using RPA technology;
• Reducing the load on the monitoring group;
• Ensuring data consistency and process transparency;
• Minimizing human error (reducing the role of the human factor in decision-making).

Automatic incident coordinator

Customer projects have a complex structure, and there can be many teams working on the project. One of the main tasks is the qualitative distribution of incidents, both between teams and within teams, when assigning incidents to a specific engineer. One of these cases was a project with 11 teams (about 50 engineers in total), where ICL Services experts were able to distribute incidents in the customer’s project with an accuracy of more than 90%.

To implement such a mechanism, the solution developers built an entire architecture containing several learning models, and all this was debugged and interconnected. In the final implementation, the program has the following functionality:

• Distribution of incidents from the shared queue to the team queue that is most suitable for incident resolution;
• Distribution of incidents from the «incorrect» queue to the correct one using intelligent evaluation
• Selection of the most competent queue engineer to solve the incident, according to the model assessment.

The development team used (in addition to the main results of the model prediction) a calculation system. This is necessary for the correct qualitative and quantitative distribution of incidents among engineers. With the calculation system, the model began to take into account not only the competence of the engineer, but also the real circumstances. For example, if an engineer is present or on vacation, or if he has an excessive or insufficient workload.

At the moment, the application is fully automatic. The program is trained daily to monitor changes in project activities and make the best decision. The software solution has left the piloting phase and is now in full production. The engineer in the role of «incident coordinator» can take a breath and do more important things, and the rest of the engineers in the team have more time to solve the incident because the coordinator does not spend this time on clarifying the circumstances.

This automation saves about 11 FTE (1800 working hours) per year, which constitutes 22% of service time. This solution is now being implemented in all major projects, and in the near future will become part of the standard for the provision of support services.

Machine Learning to support the IT Department

The introduction of artificial intelligence in the work of the IT Department entails decreasing the time for processing requests and reducing the agents’ load. At ICL Services there are two key smart projects:

• A robot for processing requests on the side of the ITSM system;
• A chatbot to communicate with users.

Gosha robot

The pilot implementation of the Gosha robot to handle requests on the side of the ITSM system was successfully completed at ICL Services in December 2018. Then there were revisions, additional training of the robot, and its functionality was supplemented. The robot was created in order to increase automation in the processing of user requests.

As a result, the robot:

• Monitors new requests in the support queue;
• Displays the «Service» and «Request Type»;
• Determines the availability of agents by their Skype for Business status;
• Assigns requests to agents;
• Transfers part of the requests to 3rd line support.

As part of the project, it is planned to train the robot to handle certain requests independently. It is intended to transfer requests that have model solution templates, where the solution does not require communicating with the applicant or collecting additional information. These capabilities of the robot will reduce the load on support agents and reduce the average time of the request in the stack. In addition, reducing the load on agents will in turn free up time for training. Firstly, this involves receiving knowledge from the 3rd support line for processing more complex requests. The fewer developers are involved in support, the more time is left for development.

Chatbot

In order to minimize costs regarding requests which have standard answers, the development team at ICL Services created a chatbot. The chatbot helps employees to solve commonly occurring problems, such as adjusting the sound in headphones or changing an outdated password. At present, this pilot project has already begun working at ICL Services. The bot sends links to existing instructions or gives the employees advice and recommendations from its knowledge base.

The development plans for chatbot capabilities are as follows:

• Collecting additional information on the problem;
• Clarifying the problem and offering a solution;
• Creating a request in the ITSM system if there is no appropriate solution;
• Saving all collected information in the request for further processing by the agent.

The bot can be accessed through Teams, after downloading it from the app store.

The introduction of all innovations has produced significant results, including improved business efficiency and quality of services, a positive impact on economic development indicators, as well as obtaining advantages over competitors. Modern innovative solutions in the field of digital technologies are better at coping with the key tasks of any business than ever before.