The capabilities of AI solutions and their use cases have been growing rapidly since the launch of ChatGPT and the likes. Yet, for most organizations utilizing these capabilities beyond individuals asking ChatGPT for input to the daily work tasks, are still just an idea on the drawing board. The technology is here, so what does require to take advantage of it?  

At VisionWillow we find that the foundation for building your future AI-driven Assistant, automating your service management requires both structure and data – more specifically, it must be the structure and data which you define. How will you know if the answers provided by your new AI Support Agent are correct? How will you know if the automated knowledge articles are up to your standard and the experience you aim to deliver? How do we know if we are succeeding? 

Whether your organization already have an enterprise strategy for AI utilization, preferred LLM’s or technology platform, our approach is both agnostic to the underlying technology and pragmatic to align with already in-place, enterprise-wide initiatives

When identifying opportunities for AI use cases within your Service Management domain, we recommend considering both short term and longer-term initiatives/use cases. A short term could be a use case realized with existing technologies. This could be to take advantage of existing AI features in your ITSM platform. A longer-term initiative may be a solution that requires advanced LLM trained on domain data, utilizing AI technology which may not be mature yet. E.g. Personal assistant AI which can help both agents and end users. We help define these use cases and the related success criteria for the MVP and adoption. This helps ensuring that the scope stays on track and the communication on progress remain consistent as there are clear relationships between the use case, requirements and what metrics are used to measure success. 

Identifying the data and finding the structure 

As the headline indicates, it all starts with data and structure. Once the use cases have been identified, we facilitate the discovery process of finding the underlying data and structure to support development and training of the AI capabilities needed for the use cases. Examples of data could tickets, tasks & workflows, knowledge articles, process- and service descriptions etc. As the current domain data and knowledge is being analyzed, we must also ask – is it good enough?  

If your organization wants to provide your end-users with their own self-service AI support agent, we will help ensure that the data reflects the expectations you want to meet. This is where the structure comes in. In order to provide the expected experience via AI, you should already be delivering it with your support specialists today. We help map the needed structure in the work we already do to create the data we will use to train the AI on. Examples of questions to consider depending on use case: 

  • Do your support agents today use a tone of voice that you are satisfied with? 
  • How good is your team at handling end-user objections in the tickets?  
  • How are your end-users perceiving the knowledge that is presented to them in self-service – Are they using it? 
  • Are there areas in our support/customer experience that should be improved? 
  • How well do you understand the wording and tone of voice of your end-users and customers? 

We ask these questions to discover any biases or shortcomings in how you work today. Thus, getting a better understanding of the structure of which our data is being created and managed.  

If you want to learn more about how VisionWillow can help you get ready for tomorrow, please contact us today.