The most powerful feature of My AskAI is its ability to turn your help or support documentation and website into a live AI customer support agent that is available 24/7 to answer user questions.
Last updated
Was this helpful?
Your AI agent will look over all the help documentation, knowledge articles and content you add, and . This means that the answer will reference at most 6 different webpages, help documents or knowledge articles, but it may only reference 1 if all pieces of text came from the same source.
Yes, when you ask a question the most relevant text is taken from across all your knowledge and help articles at once to find the pieces that can be used to answer your question. This could mean that it answers based on a single article, multiple articles, or across 6 or more separate sources.
The length of answers varies depending on:
The knowledge articles and content you have uploaded.
The complexity of your question.
In short - it depends!
The following content can be added to your AI agent to "train" it:
Websites (publicly accessible)
Documents have a 30Mb limit and the following file types can be uploaded:
xlsx
csv
For csv files, each row resides on its own line, and each element within the row has the header of its corresponding column added to it as a prefix.
docx
txt
md
rtf
tsv
pptx
mp3
mp4 (audio only)
mp2
aac
wav
flac
pcm
m4a
ogg
opus
webm
Google Drive, up to 10Mb per file
OneDrive, up to 10Mb per file
Notion
Zendesk knowledge articles
Intercom knowledge articles
Freshdesk knowledge articles
Confluence knowledge articles
Dropbox
SharePoint
SalesForce
Shopify pages, products and customer data
You can add up to 10,000 pieces of content across these connections, please contact us via chat if you need more.
While it is possible to train your AI support agent on your support tickets on historic conversations, we strongly advise against it for a few reasons:
Agent responses are highly variable and inconsistent depending on the agent's tone, experience and demeanour, this will all feed into the AI's answers
Knowledge in tickets can very quickly become dated, an answer about a feature 1, 3 or 6 months ago may no longer be relevant today so customers will get poor-quality responses that may be inaccurate.
If you have a large volume of tickets then managing this knowledge can quickly become a very difficult task as you look through the thousands of agent responses to work out which are still relevant, and which you need to update.
Ultimately, the golden knowledge source should be your help documentation and knowledge articles, as these will have been verified, reviewed, and kept up to date. Keeping them updated also means that the customers who don't use your chat still benefit from the updated knowledge.
If you want to train your AI on tickets, we recommend that you create a curated list of customer questions and agent responses as a 2 column spreadsheet (ideally added via one of our so it will update every 24 hours as you update it) and add this to your AI agent's knowledge.
We recommend that if you do want to use your historic tickets to train your AI support agent, you should use them to create knowledge articles using our features.
Your AI support agent is unable to output page numbers currently, this is in part .