How can you use chatbots to streamline IT service desk management?

The benefits of a chatbot for the Service Desk are numerous, however, there are pitfalls that management should consider. Benefits include enabling the Service Desk to "shift left" by allowing their customers to access support anytime, without delay, on multiple platforms. Second, a well-designed chatbot will address commonly occurring Service Request and Incidents and provide solutions. Finally, emerging AI Large Language Models (LLM) are delivering better interactions for customers. Pitfalls to consider, chatbots augment the human at the Service Desk but cannot replace it. Second, most customers are less forgiving of a poor experience on a chatbot. Finally, failing to monitor your chatbot will have an adverse impact to desired outcomes.

In my experience building, deploying and consuming chatbots, high volume requests/issues that take less than 15 minutes to complete are great candidates for a chatbot conversation. Ensuring all conversations provide an experience that’s different is critical. If I have to put in my name and email address and describe my issue to a bot that’s going to transfer me, so I can verify/repeat information, it’s not helpful. However, if my chatbot can do things like clear those pesky “hung sessions” in Citrix, while logging the admin function for auditing, I’m much happier not having to talk to anyone and restoring my ability to work in a self serve fashion.

Adding to what Richard said. Can we use the tools to make sure that when someone creates an incident that they put enough information in the ticket? None of us want to answer the same questions over and over, but if they are complete at the start the people who troubleshoot won't have to ask them.

Chatbots streamline IT service desk management by offering 24/7 automated support for common queries like password resets, freeing human agents for complex issues. They automate ticket classification and prioritization, reducing response times and improving resolution workflows by gathering preliminary issue information. Integrating with the organization's knowledge base, chatbots enable users to access self-help resources, reducing ticket volume. As chatbots learn from interactions, they increasingly provide accurate responses, enhancing service desk efficiency and user satisfaction. Chatbots not only boost operational efficiency but also improve user experience by providing immediate assistance and empowering users with self-service.

We can design three different types of virtual assistants. 1. Chit-chat, where users ask simple questions that don't require keeping a context, 2. QnA where users are asking questions where, eventually, we need a context to provide the correct answer, 3. Transactional, where the VA executes tasks on the user's behalf. The new large language models (LLM) can be very useful in the first two cases. Moreover, GPT technology has been proven to be less expensive than traditional approaches based on utterances, intents, and entities.

There are common issues that users confront and IT service desk management repeatedly addresses them. One way will be to analyze these common difficulties and equip chatbots with different "skills". As users struggle to complete a task, the system should identify these and chatbots should suggest a way to mend or provide few tips or refer to a helpful link. One more recommendation will be that chatbots should not pretend to know things that it can not provide answers. It should not be a data collector tool, rather a problem resolution tool. In GenAI era, it is possible to crawl thru manuals and other resources to provide near real-time pertinent responses.

Rightly said.. chatbots are more of collecting tools and would have been designed to intersect the problem solving tools .. however security is a major concern!!

Uma melhor prática muito importante ao implantar chatbots em um service desk é manter uma curadoria ativa nas intera??es. Chatbots s?o bem-vindos no atendimento quando s?o de fato inteligentes. Evitar repeti??o de perguntas, identificar as situa??es que o bot n?o está conseguindo resolver e tratá-las e, principalmente, mantê-lo sempre atualizado s?o práticas que podem apoiar e muito o sucesso do uso dos chatbots no Service Desk.

An important aspect to consider is Quality Assurance. When we delegate tasks on an automated system, we need to keep under strict observation how our Virtual Assistant performs: the quality of the dialogue, the feedback from the customer, or the tasks that are able to be solved. There are different approaches. One of the most adopted is based on human feedback. During the years, I also learned not to trust feedback but to develop KPIs to determine the VA's behavior based on data.