How KB "suggested articles" work
Question:
Answer:
Here's how Knowledge base suggestions work: it looks at the words in the subject line, then tries to find kb TAGS with this word, and suggests article with this TAG.
However in our SaaS version we also use AI - we utilize "large language models" aka "LLM's" to vectorize the content of every article and then find the closest "vector" to an incoming ticket. If you enable "ChatGPT integration" in the Settings (under "Admin - Integrations") the app will use OpenAI's GPT-model for this task. If not - the app will use our local (but much less powerful) large language model for vectorization. These articles are added to the list too.
The best way to tailor suggestions in the self-hosted version, to better fit your needs, is to "tag" articles in the knowledge base (by using helpdesk app's "tags").
I need to get a handle on how the suggested articles are pulled. I thought it was from the tags and/or the email subject
Answer:
Here's how Knowledge base suggestions work: it looks at the words in the subject line, then tries to find kb TAGS with this word, and suggests article with this TAG.
However in our SaaS version we also use AI - we utilize "large language models" aka "LLM's" to vectorize the content of every article and then find the closest "vector" to an incoming ticket. If you enable "ChatGPT integration" in the Settings (under "Admin - Integrations") the app will use OpenAI's GPT-model for this task. If not - the app will use our local (but much less powerful) large language model for vectorization. These articles are added to the list too.
The best way to tailor suggestions in the self-hosted version, to better fit your needs, is to "tag" articles in the knowledge base (by using helpdesk app's "tags").
Get help for this page
Jitbit HelpDesk