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.
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. |
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. |
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. |
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. |
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)". |
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. |
- Cursor Movement: Move your cursor over the code snippets as you mention them (e.g., hover over
LangGraphin the code). - Transition: When jumping from "Code" to "UI," use a smooth slide transition to emphasize the connection between implementation and experience.
- 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.