Artificial Intelligence – AI – is a popular topic at service management events. It’s exciting and full of promise for a brighter future. But what are the advantages of AI? How will AI help you improve your service management? Let's answer some common questions.
Artificial Intelligence is a broad concept. What you’d describe as AI some 20 years ago – think of IBM’s chess computer Deep Blue – has now become standard technology. When people talk about AI nowadays, they often mean Machine Learning. With Machine Learning, a computer uses examples to learn how to complete a task.
So the hype is really about Machine Learning, a technology with huge potential. But how will this impact our everyday lives? And how will Machine Learning change our work in service management? We’re starting to get a better idea.
Research by Gartner also shows that Machine Learning is at the peak of its hype cycle. Expectations are high; in some cases people think that the entire service desk will be replaced by intelligent chatbots within a couple of years.
We’re slowly realizing that a lot of these expectations won’t be met for now. I predict that the high expectations of Machine Learning will normalize in the coming years, and it will become more clear what the practical uses are of this technology.
Machine Learning is good at recognizing trends and changes in these trends. For service management, this means that Machine Learning quickly recognizes when an above average number of calls are logged, or lots of calls within one category, and you can act upon this.
Just imagine that you suddenly notice several calls being logged for the same problem. For instance: a specific cloud service is no longer available. Machine Learning will recognize this and send a signal to a specialized agent, who can react to this. In a more advanced setting, Machine Learning can create a major call to which other calls are linked, and which informs all agents about the call’s progress.
Machine Learning is not only able to spot urgent disruptions, but also long-term problems. Problem management is currently very labor-intensive, because you have to plough through long lists with calls. Machine Learning can help you recognize patterns in all your calls. This makes discovering structural problems a lot easier.
Machine Learning can also help you find solutions for incoming calls. At first this will only be used by service desk staff. For instance, when they receive a new call, they are shown similar, completed calls that includes details the agent can use to provide a solution.
Later, Machine Learning might be used to answer the question automatically – or even prevent calls if you provide the customer will a solution while they’re typing.
The biggest misconception is that chatbots and virtual assistants will make service desk employees redundant within the next couple of years. This won’t happen for at least 5 years.
Why not? Because the technology isn’t sophisticated enough. There are many technical possibilities – just look at the impressive Machine Learning experiments at Google and Microsoft. But the majority of chatbots for company applications are quite expensive, and don’t always work well. Asking a chatbot a question doesn’t have any added value compared to a typing a search query in Google.
Additionally, Machine Learning will mainly help to answer straightforward questions more quickly. But there are a lot of calls that are more complex, which will need to be solved by people. In fact, because Machine Learning helps solve the easiest questions, there’s more time to solve the more complex calls. And it’s those calls that are currently left too long.
Once Machine Learning becomes better at finding the right information, service desk employees will have both hands free to focus on customer experience. How do you deal with a frustrated customer? What’s the right tone you should use? What can you offer them to exceed their expectations? This requires you to act empathetically. And I can’t see Machine Learning doing that any time soon.