Feature
Generative AI can indeed have a profound impact on service desks, as it can automate responses to common incidents, predict issues based on patterns and even facilitate self-service for users — thereby freeing up human agents to handle more complex issues.
According to the Information Technology Infrastructure Library (ITIL), a service desk is “the single point of contact between the service provider and the users. A typical service desk manages incidents and service requests and also handles communication with the users.”
And while there are many places in IT that could benefit from gen AI, I wanted to hear what CIOs think the service desk potential is and where they would like service desk companies, like ServiceNow, to be investing. For ServiceNow, investing in AI could mean enhancing its platforms with capabilities, such as natural language processing (NLP) for better understanding user queries, machine learning (ML) to improve incident resolution and predictive analytics to anticipate service disruptions before they occur.
CIOs see AI investment in the service desk not only improving operational efficiency, but also contributing to better user experiences and outcomes. This could include smarter ticketing systems that automatically categorize and route issues, chatbots that provide instant responses and advanced analytics that offer insights into service patterns and help in resource planning. Overall, the potential for generative AI in transforming service desks is significant, and it aligns well with the ongoing trend of digital transformation in IT service management.
Manual Service Desk Functions
I started by asking in a recent generative AI chat at X what service desk functions remain the most manual and impacting of business success. Manhattanville College CIO, Jim Russell says, “I find that the greatest amount of effort focuses around the processing of natural language and converting that to technical or institutional language. The request for information requires clarification and specificity to data sources.”
Anthony McMahon, enterprise architect at Target State Consulting, says, “Post-call admin work is particularly valuable. It's all valuable information, as it adds context and supports root-cause analysis, but it's heavily manual. Doing it takes staff away from value-add work.” Connektedminds CEO and Principal Joanne Friedman agrees, “I would say documenting the call and translating it to actionable insight for the appropriate team are extremely manual.”
Constellation Research Analyst Dion Hinchcliffe adds, “The hardest to automate include:
- Tier II/III troubleshooting/resolution
- Management processes
- Knowledge management
- Strategy
With generative AI, some of these are easier to tackle.” Hinchcliffe continued by saying, “The human touch is certainly widely appreciated in service. But with generative AI, the automated quality will start consistently exceeding human standards. I suspect, if this becomes true, people will start assuming that the human touch, while nice, isn't as helpful as a bot.” He adds, “Any generative AI in direct service delivery roles will have to enforce strict business rules that comply with service desk policies, an IT-specific safety layer, and fine-tuning to ensure it’s focused on the right priorities.”
Running a Better Service Desk With Generative AI
So how could gen AI make service management run better? Is it about better routing of requests? Or eliminating change risk? Friedman with Connektedminds sees the opportunity “to train gen AI on the documentation from vendors or development or Dev/Ops. This could be used to create self-level management that could not only automate the routing, but create actionable insights. Meanwhile, calling functions from within the LLM means code can be modified and managed seamlessly with governance.”
Russell with Manhattanville College sees the same opportunity, saying, “Typical gen AI models leverage some machine language translation. The speed and accuracy of translating your query or request to syntax that generates the appropriate response is what we are all chasing, including the vendors.” McMahon with Target State Consulting adds, “Service desk tools range from [interactive voice response] IVR through to ticket management. Gen AI can support all of those, but know gen AI isn't a magic bullet. It needs to be factored in as part of a full strategy.”
Hinchcliffe with Constellation Research says, “Generative AI can improve service desk functions by:
- Understanding underlying ITSM databases via a holistic large language model
- Deriving strategic insights + trends
- Synthesizing new knowledge
- Automating higher order tasks
- Decision support
- Governance”
See more: How to Leverage Your AI-Powered Customer Support Strategy
Implementing Generative AI to the Service Desk
Given so much potential, how should gen AI be implemented for the service desk? By service desk tool vendors or IT or some combination? Russell with Manhattanville College is candid, “For smaller shops or colleges, we will have to wait for the vendors to deploy. But much will depend upon the quality of the knowledge base, and organizations can start that now. Given the pace of development, it will be there sooner than expected.”
Hinchcliffe with Constellation Research is right when he suggests, “The trend in larger organizations is the service desk is more than just IT. Now a dedicated central support function for HR, IT and operations. It is even deployed for the line of business. Most likely, vendors will provide most of the generative AI technology plus processes. But organizations with IP to protect must take care.”
Considering Generative AI Risks for the Service Desk
Gen AI has many business risks. How should risk not limit the potential here? Russell with Manhattanville College says “so much of the risk is the same as having untrained, unskilled or confused human operators on the desk or in a process. Ensuring you have a maturation plan for the AI models, data and quality assurance will help.”
With this, McMahon with Target State Consulting says “experiment and measure all potential use cases. Use the evidence to build the case and show where mitigants to risks are. Most of the risks are low in their nature, but without evidence, you can't justify that.” Hinchcliffe with Constellation Research agrees with McMahon and adds an industry perspective. He says, “Fortunately, generative AI in the service desk is not high risk. Risks are to service quality, [customer satisfaction] CSAT and business operations, and I'd suggest for most organizations, they are manageable. For this reason, AI for the service desk is a perfect proving ground for early gen AI [proof of concepts] POCs.”
See more: AI Enhances Customer Support, But Watch Out for These Common Pitfalls
Benefits of Generative AI in the Service Desk
The utilization of a large language model (LLM) like gen AI in a service desk environment indeed presents a low-risk and high-value opportunity. Leveraging gen AI to augment knowledge base functions can significantly enhance the efficiency and accuracy of incident, problem and request routing. Here are some points on the impact this technology can have:
- Speed: Gen AI can provide immediate responses to common queries, reducing wait times for customers.
- Learning: With machine learning capabilities, gen AI can continuously improve its understanding of how to handle different types of service desk tickets.
- Empowerment: Empowering agents with AI assistance means they can resolve issues more quickly and accurately, potentially reducing the need for escalation.
- Knowledge sharing: Turning the LLM onto vendor knowledge bases can lead to a more informed service desk where information is readily available to agents.
- Resource optimization: By automating routine queries, human agents can focus on more complex issues that require a personal touch.
In conclusion, integrating gen AI into service desk operations can be transformative, streamlining processes and providing better outcomes for both service agents and customers.
