Conversational AI in ServiceNow | Elevating Self-Service with Virtual Agent + GenAI
By: Ethan B.
Now Assist in Virtual Agent
ServiceNow’s built-in generative AI feature, Now Assist, can be used with Virtual Agent to enhance multiple aspects of employee self-service. Unlike more traditional methods of gathering user input, Now Assist is capable of understanding the context of user queries. Similarly, responses fueled by Now Assist are also more flexible than the more limited standard bot response options, and the agent can adapt its responses to be more appropriate to the current context.
The utility that Now Assist provides to Virtual Agent goes beyond its user-facing dialogues. Because Virtual Agent workflows can integrate with workflows in the Flow Designer, Now Assist can leverage the powerful tools there as well. Conversations can run flow actions or subflows to handle different self-service processes, such as creating incidents or catalog requests. A well-constructed, hybridized approach that utilizes both deterministic workflows and the probabilistic functionality of Now Assist can offer significant improvements in optimizing self-service processes, harnessing the strengths of both technologies when appropriate. Through the Flow Designer’s Integration Hub, Virtual Agent also supports integration with third-party services such as AWS, Jira, or Snowflake, allowing it to draw on data from outside of ServiceNow. These tools can be used across multiple applications, including ITSM, CSM, WSD, and more.
Outcomes and Benefits
Pairing Now Assist with the Virtual Agent results in a greater number of self-service issues resolving without the intervention of a human agent: ServiceNow’s deployment of Now Assist on their own service portal resulted in a 14% boost to employee deflection rate. Rather than speaking to a stiff chatbot, employees interact with a context-aware, more intelligent AI agent that has a better understanding of their needs, enhancing employee satisfaction with the self-service process. Now Assist’s automatic resolution note generation also reduces the time spent on that step of the resolution process substantially, with ServiceNow recording a time reduction of approximately 80% for each resolution note. Human agents are less burdened by simple issues and have greater freedom to focus on servicing employees with high-complexity issues.
In addition to benefiting employees using Virtual Agents, Now Assist’s flexibility can also reduce the costs associated with configuring them. Setting up Virtual Agents with Now Assist avoids the lengthy time costs that come with Natural Language Understanding (NLU) topic discovery. Instead, Now Assist’s LLM-fueled topic discovery only requires a topic description in plain language, rather than needing the construction of complex NLU models. With proper utilization, ServiceNow’s AI tools can improve results across every step of the self-service process, from initial configuration to use in production.
References
Performance metrics – https://www.servicenow.com/customers/now-on-now-now-assist.html.