End users may be satisfied with their level of IT service and support on the surface, but they still want to have immediate, human connections with support when things go wrong. Here is why automation isn’t always the answer, and how to affordably provide that support utilizing AI tools.

by Roy Atkinson
Date Published September 29, 2021 - Last Updated January 20, 2023

We’ve all heard the nicknames for IT and IT support: the department of no, the helpless desk, and so on. We’ve also heard that end-users and/or customers don’t want to talk to us; they’d rather fix things themselves. There are elements of truth to these statements, but they don’t tell the whole truth by a long shot.

If we mentally rewind back to the rapidly unfolding events of February and March 2020, we’ll remember the amazing feats accomplished by IT and support teams all over the world as millions of employees moved to working from home. IT provided equipment and software, rewrote policies, created guidelines, shared knowledge, moved applications and storage to the cloud, and enabled operations to continue.

A year and a half later, how do people feel? Well, according to the data compiled by HappySignals, Ltd. for their Global IT Experience Benchmark, end users are more pleased with their support experiences than ever. In fact, says the report, “End-user happiness with resolutions to IT incidents has increased over the last 1.5 years, and is now at a record high.”

But that’s only part of the story. The same report, which has a massive sample size (948,280 responses between December 2020 and May 2021), tells us that end users really like collaboration with IT, and that they like human interactions with IT more than purely technological ones. The IT portal scored lowest of the eight touchpoints tracked. People want to interact with IT, and they like to do it person-to-person.

These findings, of course, fall directly in line with the thinking that has been espoused by HDI for years. Qualities like empathy and skills like verbal communication are vital for support even in the age of machine learning and predictive analytics. People matter.

One size does not fit all

Not everyone will reach out for support every time, however. According to the benchmark study, more than half of end users are “capable of solving IT-related problems themselves.” Other segments of the user base may need a little—or a lot—of assistance when something breaks or when they have a request. How should service desks respond to these varying needs?

Support services should be tailored to the needs of the user base. Providing robust self-help is important, and so is providing assisted support for those who need it. Concentrating efforts on one at the expense of the other does a disservice to the users we are employed to support.

The myth of “call deflection”

As I wrote here in 2016: “In the never ending quest to lower costs, it has become popular to talk about deflecting calls to other less expensive channels such as self-service, chat, web forms, and even email.”

But whose costs are we lowering? If we are abandoning users who need support, they will lose productivity, sometimes with very serious consequences. Take, for example, the physician who needs to see a patient every 15 minutes to make a practice viable. Should they be required to stop treating people when they encounter an IT problem? What about a salesperson in the middle of a major deal or a grant writer facing a funding deadline? Their productivity is likely worth much more than the cost of assisted support, and failure to understand that is the result of siloed thinking. Improperly reducing the cost of support will increase costs elsewhere in the organization, impair productivity, and undermine positive business and customer outcomes.

Yes, self-help should be provided, and it should be available without driving users to a portal. Chatbots enabled by AI and machine learning can help, and there are many such tools available on the market today. Ensuring that these tools have access to good information and do not lead users down dead-end paths can serve tech-savvy users well and get them back to work fast when they encounter an issue.

Those who need assistance, meanwhile, should have multiple options for contacting support. That’s what I meant when I said in that 2016 article, “Instead [of call deflection], we should be talking about contact management, and we should have an ongoing dialog with users and customers about how they want to contact us and why.”

Live chat, for example, is somewhat less expensive than telephone and still gives users synchronous support. Remote control tools have been around for a long time and have evolved into powerful methods of live assistance.

The day is rapidly approaching when the voice channel will also commonly be AI enabled. Consider the ways in which Hey Google, Alexa, and Siri provide assistance now. Similar natural language processing (NLP) tools are arriving on the market daily for use in contact centers and technical support desks. These tools will provide ways to lower support costs without downgrading service availability, avoiding the cost “whack-a-mole” effect described above.

As you consider your support plans for 2022, I urge you to have serious conversations with your end users and customers about how they want and expect to get the support they need. Consider shifting from Service Level Agreements (SLAs) to Experience Level Agreements (XLAs) if you have not already done so. You will have a clearer picture of what is expected, and you’ll be far better equipped to provide the kinds of support your organization needs.

Tag(s): supportworld, service quality, service management, best practice


More from Roy Atkinson