According to a Zendesk study, organizations receive an average of 492 tickets per month — 69% of which could be resolved with a single touch. On average, these tickets cost $15.56 per ticket and take more than 24 hours to first respond. Based on my experience, it could take an additional 2-3 days to get the ticket resolved after the initial response has been sent.
Considering that the majority of tickets could be resolved within minutes rather than a full day, there is much room for improvement. Without improvement, there are drastic consequences for your organizations: 22 minutes of lost productivity per day per employee! And lost productivity means lost revenue.
To resolve this, you could either hire and train more support staff, create and enforce a more aggressive service level agreement, or tackle the root cause of the problem: eliminate the need to raise support tickets in the first place and look to self-service and automation to take care of a large portion (if not all) of your one-touch L1 tickets.
While we have talked about how to provide your end users with the tools to improve their device's health and stability using personalized recommendations and scorecards as well as how to proactively and reactively manage your IT environment using automation, we want to take a closer look at how much automation is appropriate before allowing your users to raise a support ticket.
How To Design An Automated Self-Service IT Support Service
Most end users prefer real-time, self-service options to picking up the phone and calling support to raise a ticket. In addition, self-service support tickets are a mere fraction of cost per ticket ($2 each) compared to tickets requiring IT support ($104 each).
To accomplish this, you will first need to do the following:
1) Find Your Top Talkers
First, identify the most common types of tickets that require only a one-touch routine response, and create a repository of fixes as your foundation of your self-service. Essentially, you will want to identify what the problems are that are affecting multiple users — your top talkers.
2) Create Your Fixes
Secondly, you have to understand what your agents are currently doing to resolve those issues manually. Are they already working with the users to carry out automation? Or are they running a script on their machine to some remote desktop, for example.
Once you have possible fixes for an issue, cascade them from the simplest non-intrusive to the most intrusive. For example, let's say you have an issue with your email and you are unable to access it through Outlook. You could simply restart Outlook. If the problem persists, you might need to clear some cache files, or even re-initialize the user’s profile or uninstall and reinstall Outlook.
3) Understand What The Problem Is
On the self-service side, we will need to have a way of understanding exactly what the problem is your end user is struggling with. There is nothing more frustrating than a bot giving you generic, unrelated, and therefore useless responses. For example, you could use Microsoft Cognitive Services and Natural Language Understanding to figure out what they are asking about.
4) Determine How Much & Which Automation You Want To Force
One of the biggest questions we always get is: "How much automation can we force before we should allow our users to raise a ticket?" While there is no clear cut answer and the exact level of automated steps depends on the unique situation, we have always found that three is a good number before users run out of patience and abort the process.
So, when a user has an issue and turns to self-service to resolve, you can suggest first the easiest, non-obtrusive solution and then work your way up two more steps, before offering to create a ticket.
5) Correlate The Problem To Relevant Answers
Now, you will need to correlate your answers with your database of fixes using automation to kick off processes that may resolve the issues right then and there. In addition, a well-organized library of knowledge base articles (including links that allow users to execute a fix with one click) is extremely helpful.
6) Automatically Raise Ticket For Unresolved Issues
If the issue isn't resolved after a certain number of steps, a support ticket is raised including the problem description from the user, all the fixes already applied, and any other information logged automatically in the background.
7) Measure, Analyze, Adjust
It is crucial to keep a close eye on your numbers, e.g.,:
- How many tickets have you been able to avoid because of this solution?
- How has your response/resolution time changed?
- How has your end user satisfaction rating improved?
- Are there any new top talkers creeping up that should be included in self-service?
By measuring, analyzing, and adjusting, you ensure that your performance is increasingly getting better. As your fix repository grows, the number of tickets avoided increases, as does your user adoption and happiness.