Skip to content

Latest commit

 

History

History
76 lines (55 loc) · 5.03 KB

File metadata and controls

76 lines (55 loc) · 5.03 KB

Slip Milestone 2: Presentation & Visual Guide

This guide maps each section of your Milestone 2 Video Script to the specific UI pages, code snippets, and diagrams available in the Slip codebase. Use this to ensure your visual narrative perfectly matches your spoken script.


📽️ Section I: The Evolution (0:00 - 0:45)

Narrative Focus: Moving from predictive ML to Agentic Intelligence.

Visual Component Resource to Show Description
Project Identity app.py Show the "Slip" logo (Slip) on the Home Page.
Landing UI app.py The Hero Shell with the tagline: "Predict churn early. Retain every customer."
Platform Stats app.py The metric cards showing Knowledge Base ✅ Ready and Vector Index ✅ Ready.

🧠 Section II: The Agent Architecture (0:45 - 1:45)

Narrative Focus: LangGraph, State Management, and the DAG.

Visual Component Resource to Show Description
The Workflow Diagram agent_workflow.md Show the Mermaid DAG (analyze -> retrieve -> generate).
Agent State agent_engine.py The AgentState TypedDict showing how context is preserved.
Graph Definition agent_engine.py The create_agent() function where nodes are compiled into a workflow.

📚 Section III: The Knowledge Engine (1:45 - 2:45)

Narrative Focus: RAG, FAISS, and Retention Playbooks.

Visual Component Resource to Show Description
The Playbook Content retention_strategies.md Show the raw Markdown file containing the Contract & Billing Strategies.
Vector Store Logic rag_utils.py The code initializing HuggingFaceEmbeddings and FAISS.save_local.
Retrieval Node agent_engine.py The retrieve_knowledge function that queries the vector store.

⚡ Section IV: Live Demo: The Strategist (2:45 - 4:15)

Narrative Focus: Interactive reasoning and personalization.

Visual Component Resource to Show Description
The Dashboard Tab UI: AI Strategist Toggle to the "AI Strategist" tab in the sidebar.
Input Section app.py Show the "Specific Retention Query" text area for custom instructions.
Thinking Process app.py The st_status block that prints the agent's thought log in real-time.
Final Strategy UI app.py The Markdown strategy output and the "Send via Email Client" button.

🛡️ Section V: Technical Reliability (4:15 - 4:45)

Narrative Focus: Fallbacks and Heuristic Mode.

Visual Component Resource to Show Description
Multi-Provider Logic agent_engine.py The get_llm_response function that iterates through Gemini, Groq, and Mistral.
Heuristic Fallback agent_engine.py The hardcoded heuristic_report used when AI providers fail.
Fallback UI app.py The UI message: "Strategy Generated (Safe Mode Fallback)".

🏁 Section VI: Closing Impact (4:45 - 5:00)

Narrative Focus: Predictive. Agentic. Intelligent.

Visual Component Resource to Show Description
Final Summary UI app.py The Product Flow summary box on the Home Page.
Full Dashboard UI: Overview Tab Switch back to the Overview tab showing all the vibrant Plotly charts.

💡 Quick Presentation Tips:

  1. Cursor Movement: Move your cursor over the code snippets as you mention them (e.g., hover over LangGraph in the code).
  2. Transition: When jumping from "Code" to "UI," use a smooth slide transition to emphasize the connection between implementation and experience.
  3. Real Data: In the Live Demo, type a query like: "Customer complained about Fiber pricing, focus on a value bundle." to show the agent actually thinking.