One of the more promising technologies at this year’s Service Desk & IT Support trade show was undoubtedly AI — but “AI” is a very wide umbrella that covers any number of different techniques and approaches. For example, AI might equally mean a chatbot that improves self-service experience for users, an intelligent knowledge base that surfaces relevant KB articles for operators to review, or better upstream filtering to reduce noise and irrelevant tickets in the help desk itself.
This wide breadth of applications, together with the sometimes over-enthusiastic claims associated with AI, have led many practitioners to be reflexively skeptical of anything presenting itself as AI — which explains why busy operations staff aren’t holding their breath waiting for it.
On top of this natural skepticism of any new hyped-up technology, there are also valid questions about how AI will fit into existing processes and job definitions. As with any extension of automation into new domains, employees worry that their jobs might be in jeopardy.
The Service Desk Future is Now
One concrete application of AI which has already shown proven benefits is AIOps, the application of algorithmic and machine learning techniques to IT operations. AIOps works to prevent the proliferation of tickets by automatically identifying significant events, and by clustering similar or duplicate events. This approach eliminates useless tickets by ensuring that only relevant and actionable issues get as far as causing a ticket to be created and submitted for human review. In addition, AIOps also avoids duplication of effort, by creating a single ticket encapsulating all the symptoms of an incident, instead of multiple tickets in different technological or organizational silos.
The result is faster detection of issues — as the symptoms are visible far sooner — and faster resolution as well, bringing all available resources to bear on the problem without the distraction of irrelevancies. All of these benefits are available without the need for creation and maintenance of complex, fragile, and brittle models, rules, or filters.
Instead, existing monitoring event sources can be analyzed automatically in real time using combinations of data science, mathematical algorithms, and machine learning, to provide genuine insight in contexts where it is most relevant.
The Results are In: Alert & Ticket Volume Drops Drastically
At SITS, Moogsoft met with IT operations and service leaders from across the globe, and discussed their plans for adopting AI, specifically to improve efficiency with their service desk operations. It was clear from our conversations that the level of enthusiasm was high, though there was some concern around what it really means to make the shift.
Change is difficult, but over time the benefits are good. Moogsoft customers that have taken the leap have told us that they were able to achieve the following results with AIOps:
- 90% alert reduction
- 62% reduction in help desk tickets
- 10x increase in operator productivity
What we’ve learned is that a key enabling factor behind these results is deep integration with existing systems and processes, ensuring seamless information exchange and avoiding the creation of new divisions within the IT Operations team.
Busy service desk administrators need not fear AI; as with any other technology, it is neither good nor bad, with results depending mainly on the questions being asked of it. AIOps is a concrete and proven response to concerns around the volume of events generated by modern IT infrastructures, coupled with the ever-increasing acceleration in the rate of change of those infrastructures.
Its flexible, collaborative, and automated approach to the incident management workflow is crucial to ensuring that IT operations can keep up with the demands of end-users for more rapid and proactive delivery of IT support.
About the author Dominic Wellington
Dominic Wellington is the Director of Strategic Architecture at Moogsoft. He has been involved in IT operations for a number of years, working in fields as diverse as SecOps, cloud computing, and data center automation.