There is no denying that ServiceNow Virtual Agent leads the competition by allowing IT leaders to have outstanding automation for ITSM compliance. Still, its complicated architecture and nature of work may be challenging for internal users and stakeholders. Businesses can face many challenges using ServiceNow Virtual Agent for service desk digital transformation or ITSM efficiency. Here is a list of cons of ServiceNow Virtual Agent.

Limited NLU or natural language understanding
ServiceNow Virtual Agent has pre-built NLU conversation templates to facilitate the most common IT, HR, and customer support interactions. However, gaining a deep understanding of emotional intelligence or empathy for human inputs and providing answers is quite challenging.
If you want to enhance ITSM interactions and provide an all-purpose conversation experience for real-time resolutions, you need additional add-ons to elevate NLU capabilities.
While Virtual Agent is a separate product, it must combine with ServiceNow Studio to boost NLU capabilities and efficiency, improve NLP understanding, and solve problems.
On the other hand, if you want to reduce MTTR to help your agents reduce response time and deliver help, you need a product called Agent Assist, which allows you to use machine learning to help deflect tickets.
Low user adoption
ServiceNow Virtual Agent and NLU can work together as chatbots only for a specific department, not all departments. For example, you need four separate chatbots if you have four different departments or function areas.
It presents a unique challenge for users. With multiple chatbots at the workplace, it is quite overwhelming for users to memorize the perfect use of a specific chatbot. Unable to realize the right chatbot for IT or HR queries, people return to old communication tools to raise calls.
As the ServiceNow Virtual Agent lacks an integrated and unified user experience, user adoption drops, impacting your ROI on ServiceNow.
Dependent on continuous data support
The self-learning capability reduces the need for continuous data ingestion into the model and keeps it updated to changing ITSM scenarios.
The burden with the ServiceNow Virtual Agent is that it cannot get smart over time by observing the feedback and action provided for problems.
Retraining and model upgrades require manual effort. You also need a data science team to feed it with new data so it can understand new content and answer unique user questions.
Insufficient custom topics
In the ITSM ecosystem, new queries can arise, and chatbots must be ready to provide real-time contextual answers. This is potential for user frustration as they struggle to find custom answers from ServiceNow Virtual Agent and get frustrated with repeated answers like “try another question or rephrase your question.”
It gives you only a few custom topics, which requires you to create continuous topics and conversation flows to meet user expectations. However, continuously building custom topics and feeding them into the model is hard.
Missing advanced analytics tools
ServiceNow Virtual Agent may have reporting features but lacks significant analytics tools to provide comprehensive insights. One must depend highly on third-party integration to derive rich insights to change the NLU models and enhance their capabilities.