There have been tremendous improvements in several areas of IT Service Management (ITSM) including departments beyond IT, by expanding the number of functional modules, workflows, and function-specific services IT provides to departments such as HR, Finance, Security, Legal, and so on. However, a key area where ITSM systems have failed to deliver is in addressing the end-user’s expectations of “self-service”. In fact, most organizations are still seeing less than 5% traction in the self-service resolution of tickets. The end-user — typically an employee of the enterprise — is at the center of any ITSM implementation, however, ITSM tooling and processes rarely lead to better end-user experiences. In most large enterprises, the end-user is unaware of ITSM systems that are in place to create, update, and monitor the status of their tickets, using mostly email, phone call, or IT forms in the employee portal as their primary mode of interaction. A large volume of tickets enter a queuing system and triage process before they are picked up by a support agent. It’s typical for the user to receive a response hours or even days later, leading to increased user frustration.
The average time to resolve an IT support ticket is 3 days, 10 hours. What if you could resolve tickets instantly?
The Holy Grail: Resolving User Tickets Autonomously
Avaamo has been helping hundreds of large enterprises with concierge virtual assistants that can be accessed via their corporate messaging, intranet portal, and voice channels to help users resolve tickets autonomously. We have been able to empower enterprises to resolve over 40% of their tickets through the self-service route. Avaamo IT Concierge Assistants are uniquely effective in solving this “last mile” challenge because we’ve looked at the entire lifecycle of an IT ticket from a user’s perspective and really focused on how we can “resolve” issues for them faster. Here are some of the key technical challenges we solved “en route” that have made this technology so effective in resolving incidents autonomously.
Machine Learning With “Small Data” is Essential for IT Support Use Cases
In the world of AI, data is the new oil. However, in the enterprise world, the challenge is usually that the datasets are much smaller and, in most cases, quite difficult to obtain due to compliance and other reasons. We have, through our vertical knowledge and subject matter experts, spent a lot of time creating vertical models that are fairly accurate despite being augmented or trained using a smaller data set that is possible for enterprises to provide. The Avaamo platform has analyzed over 100 million tickets across our global customers to make accurate predictions of the issue’s category, priority, and urgency, helping to move the tickets beyond the triage stage autonomously. Our AI system uses predictive intelligence to be able to “learn” based on past tickets and subsequent actions performed by users. Avaamo uses a patented technology that can also use data from past tickets to automatically fill various fields required in the ticket creation process for quick triage of the ticket.
Advanced NLU implementation is necessary to determine the true meaning of a support issue.
In an earlier blog post, one of our lead ML Engineers, Abhi Sharma, wrote about the gory details on why NLU for IT Support is so hard. This is also an area where we have spent time and effort working closely with enterprise customers to not just fine-tune the NLU to the Service Desk/ITSM domain but also build out the tooling necessary to configure and further train the models.
The key to unlocking employee productivity: Conversational AI that works where the user is
Most ITSM systems allow users to report their incidents via email, web portal, or call into a Service Desk. Some of them also allow for incidents to be reported via enterprise messaging apps (like MS Teams, Google Chat, Slack, Zoom, Jabber etc). However, the challenge is that these methods tend to be very asynchronous, and apart from an initial confirmation receipt, most issues don’t get a response until many hours or days later.
With the Avaamo IT Concierge our goal is to resolve these incidents at the first step before they even turn into a ticket. To do this, we’ve made the Virtual Assistant available on these channels and also made them bi-directional. This ensures that users can easily report the issue and in a large number of cases, the Assistant is able to immediately respond with either “precise” answers to their questions or conduct an intelligent exchange where the user is taken through a full troubleshooting sequence or provisioning sequence, helping to resolve over 70% of the P1 and P2 issues autonomously.
Avaamo IT Concierge in Slack provisioning an instance by applying appropriate policies
Avaamo has, via its virtual assistants, helped hundreds of enterprises leverage their investments in ITSM systems by solving the last mile of enabling “self-service” for employees to be able to not just create but also resolve their tickets instantaneously. We’d love to hear from you and discuss how we can help you in your ITMS automation journey.