Identify the potential for automation
The opportunity to automate routine processes inside businesses has been created by AI technology. In this article, we will discuss how automation happens with internal employees and IT service desk. Automation means that many internal users’ queries sent to the IT service desk can be solved not by machines, not only by humans. This simple guide will give you an overview of how to identify the routine processes at the IT service desk inside your organization.
What is the result?
The result is simple. You will get a clear overview of your current structure of requests (incidents) from users, including the following information:
- List of the most repeated (routine) queries
- Average end-to-end time for solving routine queries (the time between the creation of the incident and its resolution)
- Average time needed for resolving routine queries (how much time the service desk team spends solving a routine query)
- The average cost for solving a routine query (how much the business is paying to solve a routine query)
Identify the source of the queries and determine the time frame for which you want to make an analysis (it is better to take queries for at least six months). The source can be any incident management software like ServiceNow, Remedy or other software.
Create a template for analysis, you can create a simple Excel template. It should include the following columns: Query ID, description, creation date and time, resolution date and time, hours spent, categories (using different kinds of categories will help you manage the data, so its recommended to include them, too).
Start annotating. You can use a simple approach during the annotation and add new columns for data labeling (the first column is a noun, the second is a verb). By the end of this task, you should have a proper noun and verb for each incident description. This is high-level data labeling, and later you can create more detailed labels. Look at the example below:
Now you can start with the analysis, you can use Excel to build the analytics. The fundamental values that we should have in the report are a list of routine queries, average end-to-end time, average spend time, and the average cost of the top 50 queries.
The process can take a long time. The most time-consuming task is the data labeling but you can reuse this label data in the future; it is not just a time wasted or a one-time project. If you have label data, you can easily train machine learning models, and all future incident descriptions will be categorized automatically. You just need to monitor the labeling and modify it if needed.
People will get impressed by the result of this analysis because in many cases, they do not think about the structure of internal users’ queries. The typical percentage rate of routine queries, from all of the implemented queries, is between 15% and 35%.
Why is this important?
The more you know about the current structure of the users’ queries, the more productive and effective will be your decisions. To decide on automation, you should have a clear understanding of what you need to automate. It is not convenient to automate something that can appear once a year; there is no reason for it.
Any automation should bring value to the company, and such analysis will give you a clear understating of the potential benefits. The important thing here is that you will get the calculation of potential benefits based on your historical data, and not on average market data.