What benefit will I get if I convert a NLU topic to LLM one?

Suggy
Giga Sage

We have well defined and working NLU topics.

 

Q - Is there any benefit if I convert the NLU topics to LLM? 

3 REPLIES 3

Dr Atul G- LNG
Tera Patron

Hi @Suggy 

 

Worth reading

 

https://www.linkedin.com/posts/smitha-kaipa_nowassist-work4flow-servicenow-activity-7369761818304561...

 

 

https://www.servicenow.com/docs/r/washingtondc/servicenow-platform/virtual-agent/migrate-nlu-llm.htm...

*************************************************************************************************************
Regards
Dr. Atul G. - Learn N Grow Together
ServiceNow Techno - Functional Trainer
LinkedIn: https://www.linkedin.com/in/dratulgrover
YouTube: https://www.youtube.com/@LearnNGrowTogetherwithAtulG
Topmate: https://topmate.io/dratulgrover [ Connect for 1-1 Session]

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

Thanks for replying @Dr Atul G- LNG  but that did not answer my question.

By the way, that LinkedIn post is misleading. It says that we can have HYBRID set up, which is wrong to my knowledge.

vaishali231
Tera Guru

hey @Suggy 

Yes, there are benefits to converting NLU topics to LLM-based topics, but it depends entirely on the use case. It is not always necessary or recommended to convert everything.


1. Highly variable or free-text user inputs

If users describe issues in many different ways and your current NLU struggles with intent matching, an LLM handles this better.

Example
Users say
“I can’t access VPN”
“VPN is failing after password reset”
“Remote login not working”

LLM understands all of these without you creating many utterances.

2. Large number of similar topics

If you have many overlapping NLU topics and constant misclassification, LLMs reduce topic explosion.

NLU
Requires separate topics and training data

LLM
Understands intent from context instead of strict training phrases

3. Conversational or follow-up questions

If your virtual agent needs context awareness across messages, LLMs perform better.

Example
User
“My laptop is slow”
“Also it freezes when I open Outlook”

NLU often loses context
LLM keeps it naturally

4. Faster maintenance

With NLU you must continuously add utterances and retrain models.
With LLM you rely more on prompts and less on training data.

This is useful if your topics change often.

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