Overview
A leading U.S.-based financial technology provider, focused on simplifying payment processing for banks and credit unions, partnered with SourceFuse and AWS to accelerate its internal modernization efforts. Facing challenges with siloed knowledge, repetitive SME dependencies, and onboarding inefficiencies, the company launched Genie — a secure, GenAI-powered internal knowledge assistant powered by Amazon Q Business.
The Challenge
The client’s internal teams handle complex, multi-system operations. However, institutional knowledge was spread across unstructured content — SOPs, meeting recordings, and static documentation — making it difficult for employees (especially new hires) to access consistent, up-to-date information.
Key pain points included:
- Inaccessible or outdated operational documentation
- Time-consuming onboarding and training cycles
- Repetitive dependency on SMEs for internal queries
- Lack of a scalable knowledge search solution
The Solution
SourceFuse conducted a structured architectural assessment using ArcMod.ai, followed by the design and deployment of Genie, a secure, GenAI-powered internal assistant. Built on Amazon Q Business, Genie allows employees to query SOPs, meeting notes, and operational guides using natural language, improving efficiency and self-service access to tribal knowledge.
Sample Workflow
An employee asks: “What’s the process for onboarding a new financial partner?”
Genie responds with a summarized answer sourced from internal SOPs and links to the full guide.
Key Capabilities
- Natural language interface for accessing indexed knowledge
- Serverless architecture with API Gateway + Lambda
- RBAC and guardrails for safe, compliant use
- Real-time integration with meeting transcripts and internal documentation
The Result
- Faster access to institutional knowledge, allowing employees to retrieve SOPs, meeting summaries, and internal documentation via natural language search.
- Improved onboarding and training experience, enabling new employees to become productive faster by querying Genie instead of relying on peer support or static documents.
- Reduced internal dependency on tribal knowledge, helping break down information silos by centralizing searchable knowledge assets.
- Minimal operational overhead, achieved through a serverless, highly available AWS architecture requiring no infrastructure maintenance.
Technologies & AWS Services Used
AWS Services Used
- Amazon Q Business – Core generative AI engine for knowledge search, summarization, and chat-based interface
- AWS Lambda – Backend integration and orchestration logic
- Amazon API Gateway – Real-time communication channel between UI and backend
- Amazon CloudWatch – Logging and basic monitoring for Lambda functions
- Amazon S3 – Storage for meeting recordings, documents, and migrated video content
- Amazon Cognito (or federated SSO) – Secure authentication and RBAC
- AWS IAM – Role-based permissions and least-privilege access
- AWS Amplify – Used to build and deploy a custom React-based Genie UI aligned with the client’s internal branding and secure access controls
Technology Highlights
- Amazon Q Business native index – Used instead of Kendra or Bedrock for document ingestion and retrieval
- Custom WebSocket integration – Enables real-time communication between the Genie UI and AWS backend
- CloudFormation – Used for Infrastructure as Code to automate deployment
- RBAC & Guardrails – Implemented within Amazon Q Business to ensure compliant and safe user interaction
- Transcribe + S3 – Used to transcribe video content previously stored in SharePoint and other document management systems
- Amazon Transcribe - Used for converting meeting videos into searchable summaries
