Microsoft BOT Framework, LUIS & Q&A Integrated to SYSAID : IT Support Ai BOT Solution

- June 21, 2018

Artificial Intelligence (AI) is transforming the world, enabling people to rethink their day today activities. Where ever there is any repetitive task involving humans, an Ai bot can be implemented to automate the process. Almost all organizations have an internal IT support desk that use some sort of SR ticketing tool to keep track of the SR’s which were raised. According to our analysis, more than 40% of the SR’s raised in an organization were repetitive and creating SR’s on these applications have become a hectic task for the support teams. This is because either user would directly call the support desk or drop mails about the issue and these mails would include lengthy content, screenshots of the errors and problem definitions as well. Categorizing the SR’s has become a nightmare for the support agents.

So we figured out that all these repetitive & manual tasks can be automated by an Ai solution. Our approach to the problem can be described by a high level flow as given below

Ok, lets think of the above scenario without the Ai solution in place. Typically employees would either call the IT support desk (L1/ level one), email or IM about the problem. If its a call, the agent would refer an internal knowledge base (most of the time its a Xl sheet or a knowledge portal) and if the problem can be solved then the SR is closed. But if not then a SR has to be assigned to the next level of support (depending on the complexity L2 & L3). Similarly users can communicate with the support desk via mail and IM chat. As explained earlier, when there are more than 1000-5000 employees working in an organization you are surely to find similar problems faced repeatedly and creating SR’s in the support application & solving problems is going to be terrifying task consuming lot of time.

As shown in the flow diagram , the solution would be placed between the employees / users and the SR application. The solution would consists natural language processing and image recognition capabilities for user engagements. The natural language processing engine would contain intents, utterances and entities. It would also be integrated to a knowledge management hub where all the repetitive problems and solutions would be updated. The Ai bot solution would interact with the users via mail & IM chat clients as well.

When the user’s experience a problem they could talk to the Ai bot and get advice and solutions on them. If the Ai bot doesn’t have a solution for the problem then the Ai engine would understand it and raise an SR automatically via SISAID. The bots NLP engine would match the intents ,entities and trigger a service request according to the category and subcategories.

The above said process was previously handled by humans, but now it can be fully automated with the Ai bot solution. And the best part about the solution is that the knowledge & training can be improved daily so that the bot learns continuously on the users problem and solutions for them.

Users now don’t need to even type the problem via mail or IM chat. They could simply take a snapshot of the error screenshots / messages and upload it to the bot. The bot would use the image recognition engine and identify the content with OCR and convert it into textual format.

Using bot analytics and power BI , the organizations can also check on reporting & dashboards to understand what is the most requested information / raised problem according to various segmentation’s such as company 's branches , application, application functionality ,user groups etc. This would provide valuable insights for the organization in order to streamline the IT support and make processes more efficient.

Using voice recognition assistants such as cortana , alexa & google home, users can even interact with the Ai support bot through voice. This makes the experience and process more easy for employees.