GoML Assists CertBuddyAI in Modernizing Compliance & Audit Management

Business Problem

  • CertBuddyAI, a company dedicated to transforming Compliance & Audit Management across various industries, faced significant challenges due to manual processes and errors. 
  • The complexity of compliance processes, requiring multiple system interfaces and deep domain knowledge, could have improved efficiency and accuracy.
  • CertBuddyAI aimed to standardize and automate these processes using advanced technologies but needed expertise to build an MVP (Minimum Viable Product) to pilot in specific industries.

Explore Now

Solution

GoML proposed an intelligent audit and compliance engine powered by Generative AI solutions on the AWS technology stack. The engine would focus on automating transportation and consumer product vertical certification processes.

1.

The solution leveraged the AWS technology stack for robust and scalable infrastructure and integrated multiple data sources for comprehensive analysis.

4.

Tasks were classified within the certification lifecycle for better organization and tracking.

2.

Lyzr chatbot SDK was deployed for Natural Language Processing (NLP) to design queries and generate task lists, with responses tailored to user expertise levels (Heavy, Medium, and Light Users).

5.

GoML uses integration tools like Jira for task management.

3.

Extensive datasets were used to provide detailed summaries and regulatory references, and generated tasks were pushed to an API endpoint for task management with tools like Jira.

Architecture

  • Users and admins sign up and log in through AWS PostgreSQL RDS, with successful logins granting access to upload or view previously uploaded files and chat history.
  • Admins upload files through an API Gateway, which routes them to EC2 instances for processing and then stores them in an S3 bucket.
  • Certification data is stored in an S3 bucket, providing a centralized repository for all compliance-related documents.
  • The system uses the Lyzr SDK for Natural Language Processing (NLP) with data from the S3 bucket and integrates with Weaviate for enhanced data retrieval and semantic search capabilities.
  • Queries are processed through API Gateway and Lambda functions, which utilize the Lyzr SDK for generating task lists and summaries and reference Amazon Bedrock and Anthropic Claude V2 for advanced AI responses.
  • Generated tasks are categorized and sent back to the user with detailed summaries and regulatory references, ensuring organized tracking and efficient compliance management.
Outcomes

0%

 Automated processes reduced manual errors and improved operational efficiency

0%

The robust architecture of AWS ensured scalability for future expansions, with the system’s capacity increasing

0%

Tailored responses based on user expertise levels enhanced user experience and satisfaction

0%

Successful piloting in the transportation and consumer product verticals validated the product-market fit, with an 80% acceptance rate among initial users.