ra2 studio - Fotolia


How to improve help desk support via knowledge management

Your organization's help desk support system could be more efficient. Find out how knowledge management and natural language processing solve hidden IT woes.

If IT problems lead to frustration for everyone involved, learn how to improve help desk support with knowledge management.

Far too often, the process to solve an IT ops problem involves duplicate work, not the wrong work. A problem occurs, and the user logs it to the help desk. The help desk raises a trouble ticket, and a systems administrator, developer or other person assigned to IT support fixes the problem. The issue is closed -- for now.

Two weeks later, the same problem recurs. A different user contacts help desk and reaches a different agent, who raises a trouble ticket that goes to a different support person. They fix the problem, but in a different way. Ticket closed; job done -- again.

This approach to IT support is an issue for numerous reasons. A recurrent problem means the underlying cause was not properly dealt with the first time. The time it takes for a help desk agent to log a full ticket repeatedly drains resources. In addition, the support admin has no prior information on the problem and what caused it in the past. They cannot see what actions were carried out the last time this issue occurred and why these may not have worked or only created a temporary fix.

The issue is even worse if a problem surfaces multiple times, but in different ways -- e.g., something can be fixed at a different point on the IT platform and make things work again. Each support staff member assigned to address the issue each time has a completely different fix, which may or may not work, for the elusive problem.

When the help desk applies multiple changes to an IT platform, they increase the burden of maintaining that platform. Many organizations believe their help desk systems work effectively without realizing they experience these duplicate efforts and complications.

How to improve help desk support

Many organizations attempt to improve help desk systems by constraining how everything gets logged, but I do not recommend that route. With this method, a user contacts the help desk or fills out a ticket, which offers a set of choices. The ticket might ask, "What type of problem is this?" And then it would offer choices: hardware, platform or application.

However, users don't have a clue about the source of the issue; they just know that something prevented them from doing their work. Even the help desk agent could struggle to answer these set prompts without deep problem investigation.

An effective way to improve help desk support is to bring knowledge management to the problem, with the addition of natural language processing. NLP is a kind of AI, and help desk tools that use it can be called AIOps offerings.

With the NLP method, the person who experiences an issue files the help desk ticket, with straightforward sentences that explain what is happening. The system analyzes the text and compares it with prior entries. In many cases, the help desk software can make a strong relational match between the current problem and prior incidents. This match enables the trouble ticket to indicate what the current problem is, when and where it happened before, and what steps were taken to rectify the problem. The software can even route the ticket to the individual who resolved the problem before.

When there is not a clear pattern of problems, the knowledge management aspect of a help desk system can still provide guidance. The problem is probably similar to one or more previous issues, and the system can impart that potential relationship information to the support staff members for reference. Knowledge management of this kind provides a starting point for issue resolution that takes into account the deployment's ongoing operations, rather than starting at square one.

Companies provide such AI-enhanced capabilities in help desk tools. BMC offers the cloud-based TrueSight, which plugs into the BMC Helix Remedy service management suite. ServiceNow, which in 2018 acquired Parlo and initiated the acquisition of FriendlyData, made these moves to fold NLP services into its products.

The future of NLP

NLP has a number of other applications in IT. Developers can use AI-enhanced systems to check for software that carries out tasks they need in a feature, negating the necessity to write new code. IT administrators can identify related workloads across the entire heterogeneous platform and decide if the number of workloads can be rationalized, which reduces licensing and maintenance costs.

NLP and AI help build chatbot-based systems that can interact with users directly to help speed up and automate help desk processes.

AIOps could be bad news for help desk staff members. It avoids the manual elements of problem remediation through the use of NLP-based knowledge management, and tools can be cheaper, more predictable and safer than individual staff members. A dedicated help desk team doesn't add value to the business, but instead minimizes its losses when things go wrong.

Help desk employees should evaluate their skills and look for places to move up in IT service management if the business figures out how to improve help desk support artificially.  

Next Steps

How AI in the help desk transforms IT support

Dig Deeper on IT systems management and monitoring

Software Quality
App Architecture
Cloud Computing
Data Center