Query Regarding Intent Identification in OOTB Virtual Agent

Riya25
Tera Contributor

an hour ago

Hi Team,

 

We have a question regarding the capabilities of the out-of-the-box (OOTB) Virtual Agent.

Is there a way for the OOTB Virtual Agent to identify the user's intent and route the conversation to a specific flow based on that intent?

For example, if the user's query is related to IT, can the Virtual Agent automatically invoke a specific flow or search within a predefined set of IT knowledge articles? For queries that are not IT-related, can it continue with the default search behavior and provide responses from the general knowledge base?

Additionally, could you please help us understand the best approach for identifying the intent of a user's query? We would like to know the recommended or best-practice method for intent detection in the OOTB Virtual Agent.

We would appreciate it if you could clarify whether this type of intent-based routing is supported out of the box or requires additional configuration or customization. If it is supported, please share the recommended approach and any relevant documentation.

 

Thank you

0 Helpfuls
  • 66 Views
1 REPLY 1

vaishali231
Kilo Sage

an hour ago

Hey @Riya25 

Yes, this is possible, but the implementation approach depends on the Virtual Agent capabilities you're using.

With the out-of-the-box Virtual Agent, the standard way to identify user intent is by using Natural Language Understanding (NLU). You can create and train intents (for example, Reset Password, VPN Issue, Software Request, HR Benefits, etc.) and map them to specific Virtual Agent topics. When a user's utterance matches an intent with a sufficient confidence score, Virtual Agent automatically launches the associated topic.

For your use case, a typical approach would be:

  1. Create IT-related intents (ideally domain- or use case-specific rather than one broad "IT" intent).
  2. Associate those intents with an IT-specific Virtual Agent topic or Flow Designer flow.
  3. Within that topic, perform a Knowledge Search against the appropriate IT knowledge base or execute the required IT process.
  4. If no intent is matched, allow the conversation to fall back to the default Virtual Agent search or topic routing.

If you're using AI Search or Now Assist for Virtual Agent, you can further refine the experience by using search sources, search profiles, or search contexts to prioritize content from specific knowledge domains while still providing a general fallback experience.

Best practices include:

  1. Create focused, domain-specific intents instead of a single generic "IT" intent.
  2. Train each intent with a diverse set of real user utterances.
  3. Regularly review unmatched utterances and retrain the NLU model.
  4. Configure confidence thresholds carefully to minimize incorrect topic routing.
  5. Use topic-specific knowledge searches where possible instead of relying on a single global search.

 

*******************************************************************************************************************************

If this response helps, please mark it as Accept as Solution and Helpful.

Doing so helps others in the community and encourages me to keep contributing.

Regards

Vaishali Singh

Servicenow Developer
Linkedin - https://www.linkedin.com/in/vaishali-singh-2273361bb




0 Helpfuls