How AI is helping the help desk

Service desk chatbots and automated request routing are just the beginning. AI- and ML-driven tools will soon tap predictive analytics for better decision making in incident management, demand planning and more.

Robot Artificial Intelligence chat bot

Artificial intelligence (AI) and machine learning (ML) are turning up seemingly everywhere these days, and the IT support function is no exception. In fact, experts see AI in various forms becoming a key component of the help desk in the years to come.

“Artificial cognition will, over the next three to five years, become absolutely indispensable for any form of operations or support,” says Shannon Kalvar, research manager for IT service management and client virtualization at research firm IDC.

IT self-service is nothing new. But these days it’s becoming much more sophisticated, with chatbots and intelligent search recommendations to help guide users to the right solution. AI is expected to increasingly help IT support teams in other ways too, such as predictive analytics for incident management, demand planning, and workflow improvement.

To drive efficiency with automation, Oshkosh has transformed its service desk. Watch as Oshkosh's CIO Anupam Khare discusses how the company leverages machine learning and AI to reduce the number of incidents by 47% while increasing speed to resolution from days to minutes.

One of the biggest benefits of AI for the help desk and overall IT support function is that it can “remove the manual overhead associated with high-volume, low-value service desk activities,” says Stephen Mann, principal and content director at analyst firm “In many ways it’s similar to the more familiar IT automation of repetitive tasks that allow people to be freed up to focus on higher value-add activities.”

Early AI capabilities are already here in some IT service management (ITSM) tools, says Mann, “and I expect far more tool vendors to add AI capabilities to their existing ITSM tools this year.”

ServiceNow, Micro Focus, BMC, Symphony Summit, Ivanti, IBM, ServiceAide, and Freshworks are among the ITSM vendors who have integrated AI into their products, Mann says. Other vendors such as Astound, Spoke, and Espressive don’t make full ITSM suites but deliver AI-driven help desk or virtual assistant software.

According to IDC’s Kalvar, chatbots, knowledge curation, and incident/request routing are the three big categories of AI features and capabilities that vendors are working on today with their help desk-focused software. “We are already using artificial cognition [to] help route requests for service and service restoration to the appropriate responder, and to answer simple questions directly,” Kalvar says.

Read on for details on how artificial intelligence is helping the help desk today — and how it will in the near future.

Chatbots and virtual support agents

One area where AI is advancing is in providing “an automated 24x7 first-contact chatbot experience for users,” wrote Mann in an article for That means there is always “someone” available at the help desk, even if it’s not a person.

IT support leaders are interested in deploying chatbots, says Chris Matchett, principal research analyst for IT operations management at Gartner — even very basic ones. “In the IT service desk context, chatbots tend to do little more than apply NLP [natural language processing] through a conversational platform to carry out a targeted search of the knowledge base or other explicitly scripted actions,” he says.

Many support managers are happy to do little more than that as a way of deflecting simple and repetitious queries and requests before escalating complex interactions to a live IT support agent, Matchett says.

Some chatbots, however, can take user support farther. Virtual Support Agents (VSAs) are a type of virtual assistant that provide capabilities tailored specifically to IT support and assistance in an IT service management scenario.

“They extend chatbot capabilities by also taking action on behalf of the business consumer to do things like reset passwords, deploy software, escalate support requests, and carry out changes to restore IT services,” Matchett says. Unlike typical chatbots and virtual assistants, which require extensive customization, VSAs are pre-programmed with ITSM processes and can carry out procedural escalations of incidents.

“Many vendors claim their virtual assistants and chatbots can be used in an ITSM scenario, but very few can be considered as genuine VSAs,” Matchett says. “Both chatbots and VSAs utterly depend on either scripting or a well-established knowledge management process, although we are starting to see some VSA products that also build and update knowledge bases.” (Matchett says Gartner policy prevents him from naming specific tools.)

Request and incident prioritization and routing

AI is also driving automated categorization, prioritization, and routing of incident and service request tickets.

“We are now seeing some ITSM tool vendors offer predictive analytics that suggest or even fully populate some of the form fields in a ticket,” says Gartner’s Matchett. The tools recommend appropriate incident priority, classification, and groups to assign the ticket to.

“The human operator is able to override these suggestions, so they are not replaced but provided with AI assistance to work faster and realize less overtly obvious actions to take,” he says.

Knowledge curation

Another new capability emerging in help-desk products is AI-assisted knowledge management. This includes an intelligent search function that doesn’t just rely on specific keywords but understands context and meaning, Mann wrote in his article. It recommends solutions based on what has (and has not) worked for knowledge seekers before.

AI can also deliver intelligent autoresponders that reply to end-user emails with the most likely solutions, Mann wrote. It can even help identify gaps in the knowledge base.

AI-driven knowledge curation can help not just end users, but IT support agents as well. “Many intermediate and advanced ITSM tools utilize prior and current incidents, changes, problems, known errors, and knowledge bases to suggest potentially relevant information to IT service desk agents,” Matchett says. “This data typically appears in a dedicated area of the [user interface], such as a window docked alongside the main window.”

Gartner defines intermediate ITSM tools as those that have good ITSM capabilities and provide some basic IT operations management (ITOM) functions or integrate with intermediate third-party ITOM solutions. It defines advanced ITSM tools as having a full range of ITSM capabilities and providing broad ITOM functionality natively or integrated with advanced third-party ITOM products.

Coming soon

The next wave of AI capabilities in support software “will begin to bring more process and social analytics into the core system, along with adaptive dashboards which help an individual or a team gain awareness of the current environment,” Kalvar says. This will include some predictive analytics and modeling, so that both support staff and end users can see what the likely outcome of an action might be, he says.

In addition, AI will be able to predict and identify issues and provide automated remediation for them, Mann says. These issues include all the “noise” that comes from monitoring tools, such as alerts about the network, applications, or security-related notifications. AI “better understands what each is, thanks to pattern recognition, and only flags or attends to those that really need attention,” he says.

In the future, the focus of AI in the help desk will be on high-volume, low-value tasks that are currently being undertaken manually — with the associated costs, Mann says. It will also offer greater insight into potentially large datasets that allows for better decision making “at a level that’s simply too difficult for humans, even if they had the time,” he says.

As demands on infrastructure and operations grow, organizations will increasingly look for new opportunities to automate and provide more proactive management of their environments, Matchett says. “This becomes especially important for larger and more advanced [in terms of infrastructure and operations] organizations that can take advantage of large machine- and human-generated structured and unstructured datasets to feed into the tools,” he says.

But AI capabilities can benefit any type of business looking to automate or enhance support functions, says Mann, who predicts that small businesses might be among the first users. “As with early cloud adoption, we might see smaller companies taking more risks, even though larger companies have greater potential benefits and deeper pockets,” Mann says.

Challenges and concerns

The biggest challenge right now with AI is training and support, Kalvar says. “It takes a long time to train a system, and even more time to curate it so that it remains functional,” he says. 

An AI system “has to have clean and labeled data for learning,” Kalvar explains. “Knowledge bases have to be written. Fortunately, a lot of organizations do have knowledge bases already and increasingly the ability to reach out and use online help resources alongside the internal knowledge base.”

In terms of deploying new AI features, Kalvar says, “the actual adoption of the technologies requires the same change management skills we have developed over the last decade or so in larger enterprise transformation programs.”

Probably the number one concern related to AI and the help desk is the potential loss of IT support jobs, but Matchett says those fears are overblown. “I don’t think there is a likelihood that many, if any, IT support roles will be lost because of AI, except in cases where IT service desk staff aren’t highly trained and mostly take callers through prepared scripts — something that can be easily automated,” he says.

People who contact IT support “much prefer dealing with people than chatbots or virtual support agents,” Matchett says. “AI will change the way people work,” he adds, but it will augment them rather than replace them.

Kalvar agrees. “There is always fear of automation, but also excitement about being able to work on harder problems rather than answering the same routine questions,” he says.

Removing simpler tickets at help desks could itself be a concern, Mann says, “leaving service desk people to deal with the more complicated issues and requests.” This impacts the workload mix; there are few easy tickets if a virtual agent is successfully adopted.

It also affects the required skills and capabilities needed for people who continue to work on help desks, as well as staff morale, Mann says. It might even impact end user satisfaction with the human support staff, as they will be dealing with the more complex issues that could take time to resolve, in contrast to bot-handled requests that are resolved instantly.

Exactly how useful AI will be for IT support professionals and help desks and in what time frame depends on how ready organizations are to fully leverage these technologies, Matchett says.

“They need to have mature ITSM process and good knowledge management,” he says. “The most successful [IT support] leaders will focus not on the toys, but on the little practical steps that add real value.” Most organizations today can leverage AI to fulfill software requests, manage user identity and access, and implement natural-language querying of existing knowledge bases, he says.

What’s clear is that AI is emerging as a key component of IT strategies, and going forward the help desk and IT support will be among the beneficiaries of these new capabilities.

Copyright © 2019 IDG Communications, Inc.

Bing’s AI chatbot came to work for me. I had to fire it.
Shop Tech Products at Amazon