Paving the way with predictive intelligence in IT operations
October 27, 2020
At Accenture, our global IT organization is always looking to reinvent our internal IT operations to deliver an optimal experience for our users while streamlining our back end. Over the past year, we have been increasing the use of artificial intelligence and automation and have invested a great deal in these capabilities.
Given Accenture’s platform powered and cloud-based IT posture, we are using the ServiceNow platform for IT service management. This puts us in a position to harness the power of the platform, and this includes the use of ServiceNow predictive intelligence capabilities. The goal of predictive intelligence capabilities is to automate routine IT processes with machine learning and achieve faster resolutions.
Targeting our incident ticket process
As you can imagine, in an organization the size of Accenture with more than 500,000 employees, we handle a lot of transactions every day. Currently, this volume amounts to more than 10,000 technology issues and requests a day across our enterprise. Examples of such issues are, “I can’t log into the sales platform,” “I’m getting an error message when trying to submit my time sheet,” and “I’m working from a client site and I cannot access an application.”
Predictive intelligence in ServiceNow provides the capability to automate the issue assignment workflow. This is achieved by training and re-training the product. When trained, the product assigns an issue to the correct support team without human intervention. The prediction decision is arrived based on learnings from large volumes of historical data of past issues and the resolution action taken on those issues.
Gaining speed + accuracy
We enabled the ServiceNow Predictive Intelligence platform for the use case of the auto-assignment process. It is now operational and we are optimistic about the initial results. Predictive intelligence is improving the speed and accuracy of auto-assigning tickets. This capability is now auto-assigning 1,000 incidents a day, on average, or about 20,000 incidents a month.
And it is doing it with 80 percent accuracy. Machine learning is about re-learning. So, month after month, the consistency and accuracy of the auto-assignment capability continues to increase. We’re looking at adopting other features related to predictive intelligence, such as clustering, to help resolve common defects.
<<< Start >>>
This capability is now auto-assigning 1,000 incidents a day, on average, or about 20,000 incidents a month. And it is doing it with 80 percent accuracy.
<<< End >>>
Benefiting our service desks
These results translate into several benefits for our service desks. Auto-assignment eliminates possible human errors. We’ve reduced triage time and rerouting, thus reducing the average time to resolve incidents. In addition, auto-assignment is allowing teams to better focus on complex incidents and to support increased demand with the same capacity. It also increases the productivity of support personnel by eliminating mundane tasks. Our agents are able to focus on more value-add work as well as take advantage of new skills training.
Moving forward in the intelligence space
Our auto-assignment use case is one of several initiatives in our IT intelligent operations journey. As we move forward in this space, we’re addressing other areas of the ticket life cycle. One is by using ServiceNow Virtual Agent to provide user assistance, which we have launched. Initial topics are workstation compliance, software installation requests, laptop hardware and software requests, and multi-factor authentication requests.
Another is a pilot with the ServiceNow Similarity Framework that we kicked off with a group of our agents using the ServiceNow Agent Workspace. Agent Workspace is a command center for service desk agents that helps them respond to tickets, see contextually relevant information for issues and get recommendations for resolving them with a simplified user experience.
Ultimately, the goals of our journey are to increase the efficiency, service quality and consistency of our IT processes. We also want to reduce manual work and errors. We think artificial intelligence and automation will help get us to our destination.
