For IT help desk staff, it should no longer be a question of whether AI will transform their jobs — but when. Prepare for changes around ticket workflows, troubleshooting and more.
Artificial intelligence is often touted as the next big thing in IT. Realistically, though, AI is better viewed as augmented intelligence — it’s there to help IT staff, not replace them.
But regardless of how an organization defines AI exactly, it must consider how the technology can make it more productive and competitive. In IT, specifically, one area in which AI can be especially useful is within the help desk.
The evolution of AI in the help desk
IT service management (ITSM) addresses the need for a set of solid IT processes that revolve around a help desk. The help desk is alerted to any problems within the IT platform, issues tickets and distributes them to the appropriate entities for remedial action. Then, help desk staff sends notifications to users when the problem is fixed, and monitors and audits this ongoing string of activities.
Although ITSM help desks can operate in a nominally manual setup — where human operators initiate and complete each step — automation can speed up problem resolution and improve overall efficiency. Indeed, as IT platforms have become more complex, and help desks have grown to address such complexity, the lack of intelligence in the underlying help desk system has led to a collapse in capabilities: Help desk staff become overwhelmed by the number of issues presented to them; tickets get assigned improperly; engineers and developers lose sight of what is important; and critical issues are overlooked.
This is where AI can help. But first, data needs to be aggregated suitably.
Most IT equipment creates data, which is stored within some form of database. This has enabled admins to identify and track any issues through historical data analysis. Over time, the near-real-time analysis of such data let IT teams identify potential issues before they ever became a problem. Then, IT teams started to pool data from similar devices to forecast which devices were likely to suffer issues in the future, based on an existing problem with a single device.
Today, there is an opportunity to aggregate all data from all devices in a meaningful way, and then analyze that data to proactively identify and address IT issues. IT teams can apply AI to weed out any data that is of no, or minimal, use — a capability that’s especially beneficial, as the internet of things continues to grow. Through careful architecting of a platform, admins can filter through and minimize the amount of data they need to analyze.
After data analysis and the identification of possible issues, tickets are generated proactively and automatically. If software changes can remediate the issue, an AI system can take action and close the ticket, and then notify help desk staff what’s been done. The system can also maintain rollback points, enabling humans to undo any automated changes, as desired.
Data anonymization and cloud introduce further options to apply AI to massive data sets. Not only can an organization identify potential problems against what is happening — and what has happened — on its own IT platform, but it can compare that information with activity on many other organizations’ IT platforms. As such, AI systems are far more likely to pick up potential issues before there’s any effect on users — and remediation can occur at a more leisurely pace.
AI technology can also apply historical remediation data against new issues — for example, did a certain approach work in the past? If help desk staff applied multiple approaches previously, which one was best? An AI-based system that can update itself based on optimized solutions to a problem is far better than a simple rules-based engine that will always just try the last solution that worked, even though IT platforms might have changed.
NLP assists help desk staff
Another potential use for AI in the help desk is the application of natural language processing (NLP). Although an intelligent help desk system largely identifies issues based on data streams into the system itself, some issues are still human generated. For example, users might forget their password or not know how to carry out a certain function. However, they don’t necessarily word their issues in a standard manner: “I can’t remember my password” is different from “I can’t get onto the system.”
At both a textual and verbal level, NLP technology can filter or automate human help desk inquiries. Companies such as Unisys, Numerify (acquired by Digital.ai) and ServiceNow provide NLP capabilities either directly within their help desk tools, or as an add-on for other help desk systems.
NLP can enable shift-left approaches, where FAQs, computer-based training modules or level-one IT support staff can deal with simple problems. Then, only more complex, or infrequent, problems require escalation to the more experienced staff — and NLP can ensure these staff members only receive issues that are pertinent to their specific skill sets.
Overall, the use of AI in help desk operations will continue to rise. Companies such as BMC, ManageEngine, SolarWinds and ServiceNow rapidly apply advanced AI capabilities to their systems.
Courtesy of: Clive Longbottom