Getting started with AI in IT Service Management: 3 steps to success

Digital transformation has revolutionized many areas of modern business, but the IT Service Desk has been left behind. According to Gartner, over 80% of tickets still require manual effort from agents to understand the problem, find a solution, and provide guidance to the user.  

At the same time, IT landscapes are becoming more complex, making those problems harder to solve. Analysts have seen a 14% rise in the complexity of user-raised tickets since 2019. Unsurprisingly, the result is higher costs. Spending per agent contact is up by 27%, to an average of $18 per ticket. 

All of this makes the Service Desk ripe for digitalization and automation. Most heads of IT are aware that AI can help, but knowing where and how to start with AI isn’t always obvious.  

It’s a question we hear from mid-size European organizations all the time, so in this blog, I want to outline a 3-step approach to getting started with AI (in this case, Generative AI) in Service Management.

Step 1: Analyze your current support processes  

The key to success with AI is to find an area where it will deliver the most business value in the shortest space of time, with the least amount of business disruption. You don’t have to transform the whole Service Desk at once—it’s much better to start with a quick but valuable win.  

This includes mapping your current support processes and case lifecycle and identifying where the biggest bottlenecks and inefficiencies lie today. Here at Efecte we see most case lifecycles as having five key stages: 

• Ticket created: The user has a problem that they can’t self-solve, so they raise a ticket with the Service Desk. 
   
• Classify and route: The ticket is classified based on the type of problem that has been raised, and routed to an appropriate agent for handling. 
   
• Investigate: The agent investigates the ticket, either looking for a solution in the knowledge base, asking the user for further information, or escalating to a specialist. 
   
• Communicate: The agent communicates the solution to the user, whether by phone, email, or another channel. This may result in further rounds of investigation and communication. 
   
• Document: The ticket is closed and the resolution is documented and added to the knowledge base. 

It’s also important to analyze what kind of data you’re collecting during your support activities. Understanding your data has multiple benefits. Firstly, your chosen AI solution might need to be trained on high-quality, relevant and balanced data in order to come up with helpful answers. Analyzing your data will help you plan resources for storage and processing, and avoid any privacy or ethical issues. It will also help you identify key features for the AI model, improving its accuracy and efficiency. 

Once your support processes are clear, it’s time for step 2: 

Step 2: Identify the biggest bottlenecks and inefficiencies 

Somewhere within the case lifecycle there will be bottlenecks that slow your agents down and frustrate your users. In a typical organization, those bottlenecks might include: 

• Tickets wait in a queue for classification, slowing down the time it takes to get a suitable agent assigned. 
   
• Agents are assigned too many tickets in a day, creating stress, delaying resolutions for users, and potentially reducing the quality of agent-user communications. 
   
• Agents spend a lot of time writing emails to users, either asking for clarifications or providing solutions. Finding the right language for non-technical users can slow agents down further. 
   
• Time spent documenting ticket resolutions takes agents away from solving users’ problems. While it’s necessary for the knowledge base, it can affect the user experience. 

Identifying these bottlenecks in your own Service Desk will give you a good idea of where to focus your first AI implementation. They can be found by examining reports and by interviewing agents to collect qualitative data around their day-to-day experience. Talking to agents can also help to reassure them that AI won’t put their job at risk, but will just help them to work smarter.  

Once you’ve identified where the biggest inefficiencies and bottlenecks lie, it’s time for step 3:

Step 3: Pick a first use case that will deliver quick wins (email is a good one) 

From the bottlenecks you’ve identified, pick one area where you think the introduction of AI will deliver the most business value with the least disruption to Service Desk operations. This will become the first ‘quick win’ to prove the value of AI and get agents used to having a digital assistant.  

The area you choose will depend on your business, but here at Efecte we’ve found that email can be a really good place to start. For a start, the business value is high: in our experience, around 40% of manually-handled tickets involve time-consuming use of email, so an AI assistant can unlock efficiencies very quickly. In fact, since we implemented our own Effie AI Email assistant in our Efecte customer Service Desk, we’ve found that agents can answer 30% more emails per month.  

(Effie AI for Email is an on-demand digital assistant that helps agents compose emails faster and more effectively. It’s very quick to implement, taking only a few minutes to configure, and doesn’t require any special data or user training. Our agents found it was making their lives easier within hours—and their job satisfaction has increased too, as Effie AI Email reduces cognitive load and manual effort. If you’d like to learn more about it, you can read more here.)

AI turns Service Desk agents into problem-solving powerhouses 

Introducing AI gradually into the Service Desk means you can score quick wins in terms of efficiency and productivity without overwhelming agents with new functionality.  

We found that starting with email quickly turned our agents into problem-solving powerhouses. The GenAI feature is an optional digital assistant that agents can switch on to help them write emails faster. It auto-generates clear, relevant and human-sounding text that the agent can edit if needed. 

Our Service Desk agent Aleksi says: “I can give customers a first response a lot quicker. In some cases, Effie AI has come up with additional questions to ask the customer to help with troubleshooting. I can also multitask much better, using AI to respond to some customers while I’m finding solutions for other customers.”

Get the whitepaper: Effie AI for Agents

CIOs and heads of IT know there’s a need to increase Service Desk productivity and efficiency. As I’ve shown in this blog, AI can deliver great results fast, and can also be quick and easy to implement.

This whitepaper is intended for those interested in understanding how AI-powered virtual assistant Effie AI can transform the Service Desk. Based on Natural Language Processing and secure GenAItechnologies, Effie AI for Agents helps Service Desk teams to work faster, more productively and with higher quality, while staying in full control of the data. Download the whitepaper here.