In 2018, we firmly established our position – IT service management (ITSM) and IT service desks continue their operations despite ongoing discussions about how long they will last in the era of digital revolution. Indeed, the demand for support services is growing – the Technical Support Report and Payroll Report (Help Desk Institute) for the year 2017 indicates that 55% of technical support services reported an increase in the volume of requests over the past year.
On the other hand, many companies reported a decrease in support requests last year (15%) compared to 2016 (10%). A key factor contributing to the reduction in requests was self-service support. Nevertheless, HDI also reports that last year the cost per request rose to $25 compared to $18 in 2016. This is not what most IT services aspire to. Fortunately, analytics-based automation and machine learning can enhance processes and improve support service performance by reducing errors while increasing quality and speed. Sometimes this goes beyond human capabilities, and machine learning and analytics are the key foundation for intelligent, adaptive, and operational IT support.
This article delves deeper into how machine learning can solve many problems faced by support and ITSM related to volume and cost of requests, and how to build a faster, more automated support service that employees will enjoy using.
Effective ITSM through machine learning and analytics
My favorite definition of machine learning comes from the company :
"Machine learning teaches computers to do what is natural for humans and animals – learn from experience. Machine learning algorithms use computational methods to learn information directly from data, relying on models rather than a predetermined equation. These algorithms adaptively improve their own efficiency as the amount of samples available for learning increases."
The following capabilities are available for some ITSM tools based on machine learning technologies and big data analytics:
- Support via a bot. Virtual agents and chatbots can automatically suggest news, articles, services, and support offers from data catalogs and public inquiries. This 24/7 support in the form of suggested training programs for end users helps address issues significantly faster. Key advantages of the bot include an improved user interface and a reduced number of incoming requests.
- Smart news and notifications. These tools allow for proactive notifications to users about potential problems. Additionally, IT specialists can recommend workarounds to solve issues through personalized notifications that provide end users with relevant and useful information about problems they may encounter, along with tips on how to avoid them. Informed users will highly value proactive IT support, and the number of incoming inquiries will decrease.
- Smart search. When end users search for information or services, a context-sensitive knowledge management system can provide recommendations, articles, and links. End users typically skip some results, favoring others. These clicks and view counts are included in the weighting criteria during the re-indexing of content over time, so search capabilities are dynamically adjusted. As end users provide feedback in the form of 'like/dislike' voting, this also influences the ranking of content that they and other users can find. In terms of benefits, end users can quickly find answers and feel quite confident, while support agents have the opportunity to handle more requests and achieve a greater number of service level agreements (SLA).
- Analytics of popular topics. Here, analytical capabilities reveal patterns for structured and unstructured data sources. Information on popular topics is graphically displayed as a heatmap, with the size of segments corresponding to the frequency of specific topics or groups of keywords sought by users. Recurring incidents will be detected instantly, grouped, and resolved together. The analytics for popular topics also identifies incident clusters with a common root cause, significantly reducing the time needed to identify and address the underlying issue. Additionally, the technology can automatically generate knowledge base articles based on similar interactions or related problems. Trend analysis in any data enhances IT department activity, prevents incident recurrence, and thus improves end-user satisfaction while simultaneously reducing IT costs.
- Smart requests. End users expect that submitting a request is no more complicated than sending a tweet, namely – a short message in natural language describing the issue or request, which can be sent via email. Or even just attaching a photo of the problem and sending it from a mobile device. Smart request registration speeds up the process of creating a ticket by automatically filling in all fields based on what the end user wrote or an image scan processed using optical character recognition (OCR) software. By leveraging a dataset of observations, the technology automatically classifies and routes requests to the appropriate support agents. Agents can forward requests to various support teams and can overwrite automatically filled fields if the machine learning model was not optimal for the particular case. The system learns from new patterns, allowing it to better handle arising issues in the future. All of this means that end users can easily and quickly open requests, which leads to increased satisfaction when using work tools. This capability also reduces manual work and errors and helps cut down on resolution time and costs.
- Smart email. This tool is similar to smart requests. The end user can send an email to support and describe the issue in natural language. The support tool creates a request based on the email content and automatically replies to the end user with links to suggested solutions. End users are satisfied, as opening requests and inquiries is easy and convenient, and IT agents have less manual work.
- Smart change management. Machine learning also supports modern analytics and change management. With the frequent number of changes required by businesses today, intelligent systems can provide agents or change managers with suggestions aimed at optimizing the environment and increasing the success rate of future changes. Agents can describe the required changes in natural language, while analytical capabilities will check the content for affected configuration items. All changes are regulated, and automatic indicators notify the change manager if there are any issues with the change, such as risk, planning in an unplanned window, or a status of 'not approved.' The key advantage of smart change management is a faster return on investment with fewer configurations, settings, and, ultimately, lower monetary costs.
Ultimately, machine learning and analytics transform ITSM systems through intelligent assumptions and recommendations about application issues and change processes that help agents and IT support teams describe, diagnose, predict, and prescribe what has happened, what is happening, and what will happen. End users receive proactive, personalized, and dynamic analytical assessments and quick resolutions, much of which is done automatically, i.e., without human involvement. And as the technology learns over time, the processes only get better. It is important to note that all the intelligent features described in this article are available today.
Source: habr.com
