AWS Comprehend AI Implementation
Welcome to the AWS Comprehend AI Implementation guide. This document provides comprehensive instructions on how to set up and configure the AWS Comprehend-powered AI Assistant feature in our latest release.
Note: This document is intended for our on-premises customers who manage their own ATIPXpress application. For SaaS customers who purchase the AI Assistant, Casepoint support will enable their integration.
1.1 AWS Comprehend for SaaS Customers
SaaS customers (using an Casepoint-hosted version of the application) who purchase the AWS Comprehend integration will have their configuration completed by an Casepoint Support team member. Simply coordinate with your Project Manager (or other Casepoint contact) and we’ll handle the rest.
1.2 Implementation Methods
There are two (2) implementation methods for AWS Comprehend. Customers will select an implementation method at the time of purchase, as this feature cost is variable depending on the implementation method. The options are:
Using Casepoint AWS Account (Recommended): Customers are encouraged to implement this feature using an Casepoint-provisioned AWS account. We’ll do the work for you in setting up AWS and will provide you with a few data points to plug into the application to activate the feature.
Use Your Own AWS Account: Customers may elect to use their own AWS account and pay for their use of the service on a per-page basis.
1.3 Implementation Process
To implement the AI Assistant, you’ll complete the following activities:
AWS Account: Connect with Casepoint support to help set you up with an account we provision, or set up your own AWS account.
Application Configuration: Guidance on configuring the application settings to ensure optimal performance and integration of the AI Assistant
1.4 Prerequisites
To install the AI Assistant, you must fulfill the following prerequisites:
ATIPXpress must be upgraded to v11.7.0 or above.
The ATIPXpress API must be installed. This is a requirement for using the application in all versions v11.7.0 and above.
You must have AI Assistant license(s) and AWS Comprehend license(s) for your organization, and these licenses must be applied to your application. Contact your Casepoint Project Manager for your licenses. Once you have your updated license, follow these steps to upgrade your license.
You must have an AWS account. You’ll either use our Casepoint account, or you can use (or create) your own.
2 AWS Account
2.1 AWS with Casepoint
If you elect to use our AWS account for integrating Amazon Comprehend with our application, our support team is here to assist you. Once you decide to proceed with this option, your project manager will coordinate with our support team to provision the necessary access under our AWS organization. Our support team will create a dedicated AWS account for your organization within our organizational unit.
Once the account is set up, we will provide you with the necessary credentials, including the AWS Client ID and Secret ID. These credentials are essential for configuring your environment and enabling the Amazon Comprehend connector within our application.
Once you’ve got your Client ID and Secret ID, move to Section 3 of this document for next steps.
2.2 Use Your Own AWS Account
This section provides instructions for customers who want to use their own AWS account to integrate Amazon Comprehend with ATIPXpress. At a minimum, you are required to create an account as well as one user within the account to connect to your application.
Optionally, you may elect to create Organizational Units (OUs) to set OU-level policies, etc. This is not mandatory but can be helpful for managing your application’s hierarchy. Here’s an example hierarchy:
AWS Organization: Casepoint
o Organizational Unit: AX_ Customer_Office_1
Account: Customer’s AWS Account
• User: AX_ai_assistant user
2.2.1 Create an AWS Account
To sign up for AWS, go to AWS Sign Up and click Create an AWS Account.
Follow the on-screen instructions to complete the account creation process.
Note: You will need to provide a valid email address and payment information.
2.2.2 Set Up AWS Organizations (Optional)
Follow the steps below to create an OU:
After creating your account, sign in to the AWS Management Console.
Navigate to AWS Organizations to set up an organizational unit for your company if not already set up.
Create an organizational unit named AX_AI_Assistant (this is just a suggestion; you can use any organizational unit name).
2.2.3 Create an AWS User for Integration
Go to the IAM (Identity and Access Management) service in the AWS Management Console.
Select Users from the sidebar and click Add User.
Enter AX_ai_assistant (this is just a suggestion; you can use any name as the username). This standard name helps identify the user designated for Amazon Comprehend access within our application.
2.2.4 Assign Permissions
For the AWS user you created, attach the ComprehendReadOnly policy to allow necessary access.
Navigate to the Permissions tab.
Click Add permissions.
Select and add the ComprehendReadOnly policy.
2.2.5 Generate Access Keys
Go to the Security credentials tab for the AWS user account.
Click Create access key to generate an access key and a secret access key.
Copy and securely store these keys as they will be required for configuring the integration.
3 Application Configuration
Follow the steps below to configure your application for AWS Comprehend.
3.1 AWS Comprehend Configuration
Log in to the application as an administrator and access Administration > System Administration > AWS Comprehend Configuration:
Under Client Id, copy and paste the Access Key, generated in step 3 of Section 2.2.5 (or provided to you by Casepoint, if using the Casepoint AWS account).
Under Secret Id, copy and paste the Secret Access Key, generated in step 3 of Section 2.2.5 (or provided to you by Casepoint, if using the Casepoint AWS account).
Use the Region drop-down to select which geographic location you’d like to use for search.
Note: Selecting an option closer to your data center can help optimize speeds when using AWS Comprehend.
Click Save to save the configuration.
3.2 AI Roles and Permissions
There are three types of AI user permissions within the application:
AI Config Application Role Permission: Users with the AI Config application role permission can view and edit three AI configuration pages. Click the links below to learn more about these configuration pages:
AI User Role: Application users who are allowed to use the AI Assistant feature within document management. This is a license-based role, and all AI users must have an active license.
AI Trainer Role: A subset of the AI Users, AI Trainers are selected to feed redaction data into the Redaction Knowledge Base, which contains a repository of all redactions performed by AI Trainer to identify patterns and frequently encountered data.
3.3 Using the AI Assistant
There are two main components to the AI Assistant feature. AI Redaction Templates and applying these templates to documents in Document Management. Follow the links below to learn more about using the AI Assistant with AWS Comprehend.