Questions on Customizing Dialog GPT Behavior and Rollback Strategies

AI for Service
, ,
1

Hello team, good day!

We have a few questions regarding the behavior and configuration of Dialog GPT based on our recent proof of concept (POC):

  1. Default Behavior Observation

    • During our POC, we noticed that both Dialogs and FAQs are enabled by default.
    • Is there a way to disable these by default or selectively enable them?

  2. Intent Routing Customization

    • We want Dialog GPT to handle only specific intents.
    • For other intents, we prefer to route them through alternative implementations such as:
      • findIntent API
      • NLP properties or custom logic
    • How can we configure Dialog GPT to support this hybrid intent handling approach?

  3. Rollback Mechanisms

    • In case the results from Dialog GPT do not meet our expectations, what rollback strategies are available?
    • Can we revert to a previous configuration or disable Dialog GPT selectively?

We would appreciate any guidance or best practices on achieving this setup.

Thanks in advance!
Swathi

2

Hello Yamini,

Welcome to Kore.ai Community

We have a few questions regarding the behavior and configuration of Dialog GPT based on our recent proof of concept (POC):

Default Behavior Observation
During our POC, we noticed that both Dialogs and FAQs are enabled by default.
Is there a way to disable these by default or selectively enable them?

Answer:
Yes it is possible to disable the FAQ

Intent Routing Customization
We want Dialog GPT to handle only specific intents.
For other intents, we prefer to route them through alternative implementations such as:
findIntent API
NLP properties or custom logic
How can we configure Dialog GPT to support this hybrid intent handling approach?

Answer:
No hybride intent handling is not supported.

Rollback Mechanisms
In case the results from Dialog GPT do not meet our expectations, what rollback strategies are available?

Can we revert to a previous configuration or disable Dialog GPT selectively?
Answer:
If an appl has both NLP training and Dialog GPT enabled, precedence is given to Dialog GPT. If Dialog GPT is disabled, the bot will instead rely on its NLP training.