Documents AI

Intro

Version 5.11 of 3B Forms introduces the new AI OCR capability right in Forms. Admins can now build forms and request files to be uploaded and through natural language instructions, they can extract and modify information that would otherwise need to be inputted by hand.

Setup

Head over to Setup -> Settings -> 3B Forms General Settings and set the field 3B AI Endpoint to https://x3b-ai-76854d6fa2eb.herokuapp.com/extractImage

Usage

File Upload Component

Simply drag and drop a File question type on a form, either in a parent, child or contextual object scope. Then, simply click on "Use documents AI" and configure the options that will appear.

AI Prompt Config

Require Manual Interaction - this option will create a button that the user can click on to extract the data from the document provided. By default, the AI will scan the document as soon as it is selected

General Prompt - this is the general instruction to the AI agent. Through A/B testing, you can finetune the behaviour of the agent

Extraction Fields - these are the mappings to the form fields. Each form field can have a related prompt which will instruct the AI on what to do

Tips

• In extraction fields, be careful what the expected field type/format should be. So, if you want to extract a birthdate, you need to ask the AI to return the results in YYYY-MM-DD
• If you want the AI to check a checkbox, ask it to return the results in a "boolean true" format
• You can ask the AI to format and normalize values - e.g. ask it to normalize names, remove odd characters, spacing etc
• You can ask the AI to analyse the image and categorize it. This is especially useful if you want users to upload a specific image type

Restrictions

Although the AI is capable of reading any file type, please restrict these to images. Update the Accepted file types property to "image/*" which will enable the user to upload images only

Metering

This is a new service provided by 3B and it is exclusive to our paying customers.

Security

Data is processed off platform in a LLM container. As such, the data sent to the AI will be in transit and will be routed to Heroku (Salesforce owned PaaS). We will NOT retain the files sent over to the AI, nor will we record the results of the AI requests. However, we will collect usage data in order to analyze utilization and adoption. The data we collect includes:

- Org Id that performed the request

- Date/Time of the request

- User Id that performed the request

- File size of the request

- Analytics data such as response time, memory usage, and other technical details about the request excluding the actual request itself

Data Processing Agreement

By using this service, our clients agree to our standard DPA.