What is RAPID?
RAPID = Review & Assessment Powered by Intelligent Documentation A solution that leverages AI to streamline document review processes.
Could this be useful for checking application forms at government agencies?
Here’s the architecture from the official PDF. This might actually be useful for work. Let’s try it out right away.
Let’s Try It
(1) Sample Requirements Specification For this test, I’ll run a document check on this requirements specification. https://www.sessame.jp/workinggroup/WorkingGroup2/POT_Specification_v7.PDF
(2) Service Setup
- Enable Bedrock in the AWS Management Console.
- Use the us-west-2 region.
- Use the model
anthropic.claude-haiku-4-5-20251001-v1:0. *Note: This model’s pricing wasn’t listed on the pricing page. - Open CloudShell.
- Run the official sample script. Reference page below. https://github.com/aws-samples/review-and-assessment-powered-by-intelligent-documentation
# Set the IP to match your environment
# IP4="$(curl -4 -s ifconfig.me)/32"
IP4="<YOUR_PUBLIC_IP/32>" # 例: 203.0.113.10/32 など
wget -O - https://raw.githubusercontent.com/aws-samples/review-and-assessment-powered-by-intelligent-documentation/main/bin.sh \
| bash -s -- \
--ipv4-ranges "[\"$IP4\"]" \
--cognito-self-signup false \
--auto-migrate false \
--bedrock-region us-west-2 \
--document-model "us.anthropic.claude-haiku-4-5-20251001-v1:0" \
--image-model "us.anthropic.claude-haiku-4-5-20251001-v1:0"
~ $ IP4="<YOUR_PUBLIC_IP/32>"
~ $
~ $ wget -O - https://raw.githubusercontent.com/aws-samples/review-and-assessment-powered-by-intelligent-documentation/main/bin.sh \
> | bash -s -- \
> --ipv4-ranges "[\"$IP4\"]" \
> --cognito-self-signup false \
> --auto-migrate false \
> --bedrock-region us-west-2 \
> --document-model "us.anthropic.claude-haiku-4-5-20251001-v1:0" \
> --image-model "us.anthropic.claude-haiku-4-5-20251001-v1:0"
--2025-11-01 18:04:46-- https://raw.githubusercontent.com/aws-samples/...
...
---------------------------------------------------------------------------
Deployment completed successfully!
---------------------------------------------------------------------------
Frontend URL: <https://xxxxxxxxxxxxxxxx.cloudfront.net>
---------------------------------------------------------------------------
(3) Setting Environment Variables
EMAIL="your_mail@your_domain"
# User Pool ID (the one where OutputKey contains "UserPoolId")
POOL_ID=$(aws cloudformation describe-stacks \
--region us-west-2 --stack-name RapidStack \
--query "Stacks[0].Outputs[?contains(OutputKey, 'UserPoolId')].OutputValue" \
--output text)
echo "POOL_ID=$POOL_ID"
# User Pool Client ID (the one containing "ClientId")
CLIENT_ID=$(aws cloudformation describe-stacks \
--region us-west-2 --stack-name RapidStack \
--query "Stacks[0].Outputs[?contains(OutputKey, 'ClientId')].OutputValue" \
--output text)
echo "CLIENT_ID=$CLIENT_ID"
# (If you want us-west-2 as the default for this session only)
export AWS_DEFAULT_REGION=us-west-2
(4) Creating Users
# 1) Create an admin user
aws cognito-idp admin-create-user \
--user-pool-id "$POOL_ID" \
--username "$EMAIL" \
--temporary-password '<TEMP_PASSWORD>' \
--message-action SUPPRESS \
--user-attributes Name=email,Value="$EMAIL" Name=email_verified,Value=true
# 2) Change the temporary password to a permanent one
aws cognito-idp admin-set-user-password \
--user-pool-id "$POOL_ID" \
--username "$EMAIL" \
--password '<PERMANENT_PASSWORD>' \
--permanent
(5) Login Connect to the Frontend URL output during the deployment, and successfully log in from the login screen.
Create a new checklist.
(6) Problem Encountered – To Be Continued The checklist verification never finishes.
Is this just taking a while and will eventually finish? Or is the analysis failing internally?
Rather than relying solely on LLM-based checks, maybe strict validations like numerical calculations should be handled by Lambda? I’d like to verify the whole flow of combining AI with non-AI logic for the review process.
11/2 13:11 Update
The checklist still hasn’t finished after half a day, so I checked the status in Step Functions. It looks like there was an issue with the model specification during deployment.
--document-model "us.anthropic.claude-haiku-4-5-20251001-v1:0" \
--image-model "us.anthropic.claude-haiku-4-5-20251001-v1:0"
When no model is specified, it appears to default to Claude Sonnet 4 (2025-05-14). Let me retry with the default model settings.
Re-run Results
~ $ IP4="<YOUR_PUBLIC_IP/32>" #実際のIPに合わせて変更する
~ $
~ $ wget -O - https://raw.githubusercontent.com/aws-samples/review-and-assessment-powered-by-intelligent-documentation/main/bin.sh \
> | bash -s -- \
> --ipv4-ranges "[\"$IP4\"]" \
> --cognito-self-signup false \
> --auto-migrate false \
> --bedrock-region us-west-2
...
===========================================================================
Deployment completed successfully!
---------------------------------------------------------------------------
Frontend URL: https://xxxxxxxxxxxxxxxx.cloudfront.net
===========================================================================
I fed it this file (POT_v7_checklist_from_csv.pdf) and created the checklist again… Success!
Looking at the completed checklist, some items are flagged as needing revision. I’ll go through and fix them according to the suggestions.
Created the review data. Waiting for results.
Finished in about 10 minutes.
Some items failed, but most passed. $3.67 per document feels pretty expensive, but I suppose it’s cheap compared to labor costs?
Alright, next up I’ll work on chaining in custom validation logic into the workflow.