Skip to main content

Knowledge for your AI phone agent

Add content about your business to make your agent an absolute genius

Maurizio avatar
Written by Maurizio
Updated over a week ago

What is Ringly.io's knowledge base?

The knowledge base is where you add information about your business. When a customer asks a question, the AI phone agent always checks the knowledge base for the solution.


Where can I find the knowledge base?

In the menu on the left of your screen, click 'Knowledge'.


What are the limits of the knowledge base?

Here is what you can add to 1 knowledge base:

  • 300 webpages (urls), which are auto-synced every 24 hours

  • 1,000,000 characters of plain text in total (~200,000 words) or 50 text files, whichever you reach first.

  • 25 documents (each one being max 50MB)

If you need more than 1 knowledge base, you can buy more at $9/mth each. This way you can add multiple knowledge bases to one agent.


Set up the knowledge base in 3 steps:

Step 1) Click on one of these options (more about the 3 below):

To upload content, you need to:

  • Be on at least a $49/mth paid plan to get 1 knowledge base for free

  • OR pay $9/mth for 1 knowledge base

After adding the website, text, or file, it shows up in the knowledge base with the status "Syncing..." like here:

After 10-30 seconds the file will be synced. Bigger files = wait longer:

Step 2) Add your agent(s) to the knowledge base.

You can do this two ways:

Way 1: Click on the 'edit' button in the top-right of your knowledge base

And then choose what agents will have the knowledge

Way 2: Go to your agent and click on 'Knowledge'

Then click on '+ Connect a knowledge base' and choose your knowledge base. Now it is connected.

Here are the available content types to upload:

Upload a document

Supported formats: You can upload documents in PDF, CSV, HTML, RTF or XML (max. 50MB per document)

Steps:

  1. Click on 'Documents'.

  2. Give your document a name

  3. Click on the upload area or drag and drop a file

  4. Click 'Attach file'

Upload text

Paste or type plain text.

Steps:

  1. Click 'Plain text'.

  2. Give your text file a name

  3. Paste or type your text content into the input field

  4. Click 'Attach text'.

Add a website

Use case: Allows the AI to fetch information from your website.

Steps:

  1. Click 'Websites'.

  2. Enter the URL of the website and click 'Upload website'

Wait 30 seconds or so...

3. Here are all the pages of your website! Uncheck urls you don't want and then click "Confirm and upload".

Notes:

  • If you add a url like ringly.io we will also scrape help.ringly.io or docs.ringly.io

  • if you add https://www.ecomstore/products we will scrape all urls of your website that have the /products path.

  • If the scraper did not find all your urls it can be because 1) Some pages have no links linked to them. 2) Some pages may have blocked scraping or robot access at all.


Knowledge base best practices

  • Images will use up a lot of storage space, so it's best to avoid using documents with images or styling.

  • Give your files descriptive names. This way the AI phone agent knows where to look.

  • Call your agent to see if it gives the right answers

  • If your agent makes up an answer: add to the agent’s instructions that it should only get information from the instructions or knowledge base, and if the information is not there, mention that you do not know the answer, instead of making something up.

  • Use markdown format for the source, with clear paragraphs: it will be easier to chunk and is easier to be retrieved.

  • For tabular data or image data, it’s sometimes hard for the system to recognize the correct data and put relevant information together in one chunk, so it might yield suboptimal results.

  • Group relevant information together in one chunk: it will improve the retrieval accuracy.

  • Avoid ambiguity and use specificity in references: when the chunk is retrieved, it’s best to contain specific references like name, dates, and less ambiguous terms like it or this. Because without the prior chunk, it’s hard to understand ambiguous terms.

Did this answer your question?