Current Status: 🚧 Pilot Phase
The Email Agent is a pioneering Retrieval-Augmented Generation (RAG) AI system designed to automatically reply to support emails at Brigham Young University-Idaho. This system addresses critical challenges in volume and quality of email responses to support emails while maintaining BYU-Idaho's values of Christ-like service and academic excellence.
The BYU‑Idaho Support Center (BSC) currently contends with an overflowing email inbox—far more messages arrive each day than its team of 25 part‑time student agents can feasibly address. This backlog leads to delayed responses and, as recent CSAT surveys confirm, diminished satisfaction characterized by long wait times and inconsistent answers. The manual effort spent triaging and drafting repetitive replies also diverts valuable staff hours away from higher‑value engagements such as real‑time phone or chat support.
Key Challenges:
- Low email satisfaction scores due to delayed response times
- High operational costs from managing 25+ part-time positions (estimated > $100,000/year)
- Inconsistent information delivery during peak support periods
- Training overhead and scheduling complexities for student workers
- Variable service quality impacting student experience
An intelligent AI agent that:
- Automatically processes incoming student support emails
- Searches knowledge base using semantic search upon the existing BSC knowledge base (from Genesys Cloud, vectorized with OpenAI's embedding model, stored in Pinecone)
- Generates contextual responses grounded in official BYU-Idaho policies
- Maintains human oversight through approval workflows
- Ensures compliance with FERPA and institutional values
graph TD
A[Email Input] --> B[Data Processing]
B --> C[AI Agent Core]
C --> D[Knowledge Base Search]
C --> E[Academic Calendar]
C --> F[Web Search]
C --> G[Response Generation]
G --> H[Human Review]
H --> I[Response Delivery]
J[PostgreSQL] --> K[Conversation Logging]
J --> L[Evaluation Metrics]
| Component | Technology | Source | Status |
|---|---|---|---|
| Agent Platform | n8n | https://byui.app.n8n.cloud/ |
🟡 Pilot |
| Language Model | OpenAI GPT-4.1 |
https://platform.openai.com/ |
✅ Active |
| Embedding Model | OpenAI text-embedding-3-large |
https://platform.openai.com/ |
✅ Active |
| Vector Database | Pinecone | https://app.pinecone.io/ |
🟡 Pilot |
| Database | Azure PostgreSQL Flexible Server | https://portal.azure.com/ |
✅ Active |
| Knowledge Base | Genesys Cloud | https://apps.usw2.pure.cloud/ |
✅ Active |
- n8n Flow:
Email Agent - n8n Flow ID:
8Fi9AUexVzJze7tl - OpenAI Platform:
BSC Agent - Pinecone:
AI Agents(index:bsc-knowledge-v2)
- Multi-tool Integration: Knowledge base, academic calendar, web search, escalation
- Contextual Memory: Context is preserved, email by email
- Source Citations: Sources retrieved from the knowledge base for response generation are cited in every message response
- Technical architecture design
- n8n workflow prototype development
- Multi-tool integration (knowledge base, calendar, web search)
- PostgreSQL data management system
- Values-aligned system prompt engineering
- Automated learning loop for BSC Agent (n8n, Pinecone, Genesys)
- Human review workflow implementation
- Comprehensive evaluation framework
- Stakeholder validation and feedback incorporation
- Enterprise tool licensing acquisition (n8n, Pinecone, Label Studio)
- Production environment setup
- Security and compliance review
- Production deployment planning
- Staff training and change management
- Performance monitoring dashboard
- Continuous improvement framework
- n8n workflow platform access
- OpenAI API credentials
- Pinecone vector database account
- Azure PostgreSQL Flexible Server
- Response Time: Reduction in the time required to respond to end-users
- Consistency: Standardized, knowledge search
- Cost Efficiency: Reduced reliance on manual knowledge search
- Scalability: Handle peak periods without service degradation
- Values Integration: Christ-like service embedded in every interaction
- Mission Support: Advancing BYU-Idaho's educational mission
- Student Success: Improved support experience and satisfaction
| Role | Name | Responsibility |
|---|---|---|
| Project Sponsor | Karl Karstad | BSC Director, Primary Stakeholder |
| Project Manager | Brian Schow | Project Coordination, Stakeholder |
| AI Engineer | Ron Vallejo | Technical Development, Architecture, Implementation |
| AI Governance | Sid Palmer | Governance, Oversight, and Compliance |
- CES Data Privacy & AI Governance
- BYU-Idaho Gen AI Council
- BYU-Idaho IT Project & Portfolio Management Council
This project represents BYU-Idaho's entry into institutional AI deployment, serving as:
- Proof of Concept for AI in university operations
- Learning Laboratory for AI governance and best practices
- Foundation for future AI initiatives across campus
- Case Study for values-aligned AI development in higher education
Developer: Ron Vallejo, AI Engineer
Email: vallejor@byui.edu
Stakeholders:
- Brian Schow, Project Manager
- Karl Karstad, BSC Director
This project contains proprietary code and is not licensed for reuse. The codebase is therefore gated and stored in private repositories. For compliance, this project operates under BYU-Idaho's institutional policies and Church Educational System guidelines. All AI implementations must comply with:
- FERPA privacy requirements
- CES data governance policies
- BYU-Idaho academic integrity standards
- Institutional values and mission alignment
This document represents BYU-Idaho's commitment to responsible, transparent, and ethical AI development and deployment in support of the university's mission and core strategies.
Built with ❤️ by BYU‑Idaho AI Engineering