Virtual Agent looping to Greeting topic when intent not found
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎11-16-2023 04:45 AM
We've noticed some strange behaviour in Virtual Agent when an utterance is entered that there is no Intent for; it loops straight back to the beginning of the greeting topic:
The below screenshot shows the example utterance "Help editing videos", which when inputted does not match on any of our intents:
As you can see, when I test this utterance in the designer no intents are found:
I understand this could be resolved by adding the utterance to one of our intents however, should the expected behaviour not be that the fallback topic kicks in? I have double checked our setup topics and have confirmed that a fallback topic is in place.
Many thanks in advance.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎09-03-2024 11:12 AM
@pavithrajrs Sorry for the huge delay here! I only just found the root cause of the issue. So the problem was that the description for our greeting topic was triggering a keyword match to certain utterances, which is basically the NLU's last ditch effort to find a topic to route to. You can prevent that from happening by finding your topic in the sys_cs_topic table and setting the "Is Topic Discoverable" field to false. Alternatively, if you don't want to make it undiscoverable, you can change the description of the topic so that it doesn't trigger keyword matches with any utterances. So for example, if your topic description is "does xyz" and it is getting triggered by the utterance "need help with xyz", change the description to "does abc" (or just delete it entirely).
If you haven't found a solution already, I hope this helps!
Sean
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎09-03-2024 12:11 PM - edited ‎09-03-2024 12:12 PM
Thank you so much for your reply. Could you please help me understand the steps you followed to trigger the fallback when the given utterance doesn't match any of the topics? In my case, it's giving the default message from the system properties instead of picking it from my customized fallback topic. And sometimes random topics its hitting not sure what was the mistake in my configuration,
Here are the steps I followed:
- Created a default fallback topic.
- Mapped the fallback topic in the Conversation Interface Settings.
Despite this, I'm still seeing the default system message.
In sys_cs_topic table I can see this default fallback topic has value for "Is Topic Discoverable" as true. Is that fine? Do I need to change anything in the sys_cs_topic table.
Any guidance you can provide would be greatly appreciated.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎09-05-2024 04:07 PM
Hmmm, the only thing I can think of is to try setting multiple setup topics to your "Default Fallback" topic. For example, try setting "AI Search Fallback", "Fallback", "Explore Help", "Live Agent" and "Error" setup topics to invoke your custom topic. Or try them one at a time to see if one works. That should cover most of the bases so that your custom "Default Fallback" topic is invoked when the NLU doesn't know what else to do. It shouldn't matter if your default fallback topic is discoverable, unless it is being triggered erroneously by certain utterances.