In my last post, I discussed how AI and machine learning are applied in IT. In this post, I’ll share examples that are working today.
We already reviewed how historical data is used to train machine learning models to provide better self-service. These models are used in a variety of domains to proactively schedule service calls, prevent service outages, and measure the health of IT services.
Other ways AI-driven automation delivers value include:
- Accurately routing incidents to human IT agents based on criteria such as incident category, assignment group, priority and staff availability.
- Surfacing and embedding relevant information such as related service requests, incidents, changes and knowledge into incident records.
- Enabling virtual agents to make personalized recommendations and/or triggering workflows to fulfill requests.
Additionally, advanced machine learning-based AI systems can be trained and configured to leverage internal and external data sets to not only answer specific questions but also self-learn and continuously improve responses.
Getting the right mix:
A narrow/domain focused AI system that leverages a finely tuned mix of supervised and unsupervised machine learning provides IT organizations with valuable insights into employee service and support experiences.
As a result, IT automations become more intelligent and dynamic, based on employee preferences and needs. This allows IT to better comply with SLAs and improve operational metrics, such as improving customer satisfaction scores and lowering cost per ticket and mean time to resolution (MTTR).
Real world application:
The sections below describe how an IT department within a large enterprise company is gaining tangible benefits by leveraging an AI and machine learning platform to automate the entire IT service delivery lifecycle, end to end.
In terms of automating incident management processes, the IT department is using virtual agent technology to dramatically reduce the need for level 1 support agents to perform the initial triage of service desk tickets. The virtual agent allows employees to request and/or provision IT goods and services 24/7, regardless of location. For instance, employees use the virtual agent to quickly reset passwords and request new devices and software. By shifting these types of common inquiries left, IT has been able to reduce call volume by up to 70%.
In cases where the virtual agent must involve a human agent, machine learning capabilities are used to, not only automate the escalation process (i.e., ensuring tickets are routed to the correct support agents and contain the appropriate categories and priority levels), but also to enrich tickets with relevant data such as CMDB information, app/system configuration details, related incidents and chat logs as well as standard fixes for common issues. By providing IT staff with real-time access to this type of valuable information (which would otherwise be buried in disparate knowledge bases and/or siloed across the organization), the IT department has significantly lowered its MTTR. As a result, the organization’s cost per ticket has decreased from ~$25 to ~$1.50.
Rounding out the service delivery lifecycle, the IT service desk manager utilizes operational insights to gain real-time visibility into the support team’s productivity and the health of IT services based on historical incidents, requests and remediation details. What’s more, to ensure that all processes have been executed successfully, both the manager and IT support staff rely on trend analytics for insights into how incidents are progressing based on topic, geography, team and role.
The IT department also leverages analytics dashboards to proactively assess how AI-driven automation is adding value in terms of advancing the adoption of employee self-service, reducing support costs and improving customer service.
By embracing a holistic approach to AI-driven automation, this forward-thinking IT department has decreased MTTR and cost per ticket, while increasing customer satisfaction.
Machine learning also drives many of the natural language processing (NLP) & natural language understanding (NLU) capabilities found in high performing virtual agent technology.
In part three of this series, learn about the role of NLP and NLU in AI for ITSM.
Latest posts by Robert Young (see all)
- Virtual Agents – Is the IT Service Desk Ready for the Future of Work? - March 8, 2018
- Applying AI Technology to ITSM: The Difference Between NLP & NLU (Part Three) - February 20, 2018
- Applying AI Technology to ITSM: A Holistic Approach – Part Two - January 23, 2018