ITSM chatbots: 6 AI-based use cases for your service desk

Chatbots.

Chatbots can be trained to handle a particular category of requests and incidents, provided there is proper documentation of the past request history and all relevant knowledge articles. Here, we’ll discuss two scenarios in which chatbots could help service desks: The first is an application of artificial narrow intelligence, which is available now, and the second is based on artificial general intelligence, which is more efficient but might take longer to develop.

Chatbots – Scenario 1: Resolving printer problem (AI, narrow)

One issue that seems to haunt both end users and IT technicians alike is when the printer stops working. In most IT service desks, the solution to every printer issue is already well-documented, which means many end users can resolve these issues themselves without involving an IT technician. But still, there are a number of printer incidents that get reported, which might impede productivity. Such incidents can be handled by a chatbot trained to specifically handle printer issues.

In a typical conversation between a chatbot and an end user reporting a printer problem, the chatbot responds to the user based on the available knowledge base articles. The chatbot suggests the solution with the highest success rate to the user first, followed by the other available solutions in the order of their success percentage. When the chatbot runs out of solutions to suggest, it can pull in a human technician to help the end user; it can even be trained to create a ticket on behalf of the user and have it assigned to the right technician or a support group based on past data. With multiple third-party chatbot tools in the market as well as various solutions provided by IT service desk vendors, service desks can implement chatbots right now.

Chatbots – Scenario 2: Resolving printer problem (AI, general).

As the technology driving AI advances, chatbots will be able to do more than just suggest solutions. For example, imagine the same printer issue as before with a more involved chatbot. AI algorithms and chatbots can become more intelligent than they are today, and soon, they may be able to proactively identify issues and provide the necessary resolution.

For instance, a request could be created for toner replacement even before the user reports the issue. With machine learning (ML)-based models, service requests can be automatically created for replacing toner and other supplies before they run out. And, when the user reports the issue, the chatbot could look into the requests database to determine if a request had already been created for the same issue before checking the solutions module. If and when it identifies a request, the chatbot could provide all relevant details to update the user. Although this functionality doesn’t yet exist, it may not be long before it does.

Apart from these two scenarios, there are multiple ways chatbots can come in handy. Below are a few examples.

Chatbots – Scenario 3: Remote user asset request

An end user in the field (e.g., a sales person) reports that their laptop is slow and needs to be replaced. They try to find the right asset upgrade form but can’t. They next try calling the service desk but don’t get through to anyone. As a last resort, they reach out to the chatbot.

Chatbots – Scenario 4: Add notes, comments or annotations to a request

An IT tech is working remotely to diagnose an issue with a workstation, so they aren’t able to access the service desk portal to update the request details. Instead, they use the tech assistant chatbot to get things done.

Knowledge Management

AI algorithms and chatbots are only as effective as their available knowledge base. Fortunately for us, AI can also help build a sturdy knowledge base. We will discuss two use cases to understand how AI can contribute to Knowledge Management in IT service desks.

Knowledge Management – scenario 1: Automatically rating solutions to approve and reject them

For each incident or incident category, there may be multiple solutions and knowledge base articles that have been used over a period time. Specific ML-based models can be trained to identify the success rate for each of these solutions based on historic performance. This can be done by considering multiple factors such as the reopen rate of tickets, end user and technician rating of articles, and acknowledgement from end users.

Based on metrics like these, an ML-based model can even suggest which articles should be retired and which articles could be improved. The grading of solutions based on their performance over time also helps the IT service desk provide the right solutions to users at the time of ticket creation and assists chatbots during a chat session.

Knowledge Management – scenario 2: Identifying problem areas and collating knowledge base articles

ML-based models like those discussed in scenario 1 can be trained to identify the incident categories that have the highest number of incoming L1 incidents, repeat incidents, and reopened incidents. Consequently, an ML model can flag these categories based on the severity of the above parameters. It can also deliver insight on which categories need more knowledge management efforts from the IT service desk team, such as documenting proper solutions, getting them reviewed, and publishing them. This helps the service desk team identify areas that need the most work and build relevant solutions and knowledge base articles to help end users and technicians alike.

Service Request Management

Today, complex service requests like employee onboarding are either manually coordinated by technicians or based on predefined automations. Manually performing these tasks can be inefficient and cumbersome. With respect to current automations, most processes are static and lack intelligence. These automations don’t necessarily fit all the possible scenarios and require human intervention periodically to stay on course. But with the application of AI technology like Machine Learning, models and algorithms can be trained to dynamically automate service request workflows based on request history. These ML-based automation models continue to learn with each bit of live data to fine-tune workflows for higher efficiency.

IT Change Management

IT Change Management is one process that can make or break a company’s IT infrastructure. A lot of planning and risk evaluation go into changes before they’re implemented; despite all this effort, changes can still fail due to human error. When it comes to analyzing changes, humans can also struggle to mine insights from the huge volume of data generated on IT change management and implementation. AI can help minimize change management risks by preventing human error and improving analysis.

IT Asset Management

IT Asset Management and a configuration management database (CMDB) serve as the foundation on which every ITSM process operates. AI can also help IT service desk teams monitor and manage IT hardware and software assets better. ML systems can constantly monitor the performance of a configuration item (CI) or go over the available CI performance data and predict breakdowns, saving both end users and IT teams from a pile of trouble. AI can help IT service desk tools flag anomalies and generate critical warnings by connecting the dots across multiple areas, which is almost impossible to do manually.

These are certain areas where AI will start leaving an impression on ITSM. Some AI capabilities are achievable immediately, and some are still a few years out. Chatbots and ML-based categorization will be the first immediate application of AI in ITSM tools. Some ITSM tool vendors have already started offering both these capabilities to their end users. There are also multiple third-party vendors who provide plug-and-play solutions that can perform these operations. In just a few short years, we may see more than just projected use cases for AI-based features. Soon, these use cases could be our reality.

Getting Ready for the AI Wave in ITSM

Given that AI has the potential to redefine the way IT service desks and IT service desk teams work, it’s essential that service desks are prepared for the upcoming AI wave. As explained above, the effectiveness of any AI application or model depends on the data it’s trained on as well as the available knowledge from things like documented solutions.

To get the most out of AI, IT service desk teams have to properly document all their requests, problems and changes; maintain an accurate IT service desk database; and build a well-equipped knowledge base. As ITSM tool vendors try to integrate AI into their products, it’s important that service desk teams prepare themselves to truly reap the benefits of AI in ITSM.

This article was originally published in DATAVERSITY.