Here's a succinct story map for both approaches:

---

## Data & Insights AI Agent — Story Map

### Approach A: Gradual Integration

**User Entry**
→ Arrives in existing Reports/Dashboards area
→ Floating agent UI (Jack's) is visible and contextually tied to current report

**Query & Answer**
→ User asks a natural language question (e.g. "How many family tickets did I sell last week?")
→ AI clarifies ambiguity if needed ("Across all events or just active ones?")
→ Answer returned inline — filtered existing report, new visualisation, or direct answer

**Drill Down & Save**
→ User can drill into a dedicated report view for more detail
→ Floating agent persists — ready for follow-up queries
→ User can pin/save AI-generated reports or dashboards
→ AI-generated content visually differentiated from human-created content

**Revisit & Iterate**
→ Recent chat UI lets users jump back into previous query threads by topic/job-to-be-done
→ Agent begins suggesting additional visualisations based on question patterns (Phase 2)
→ Proactive anomaly surfacing and role-based templates introduced (Phase 3)

---

### Approach B: Conversational-First Redesign

**User Entry**
→ Arrives at a new conversational surface — chat is the primary UI
→ No blank canvas — role-based templates pre-populate the space
→ Jamie-style onboarding prompts user to describe their role and priorities

**Query & Explore**
→ User queries in natural language across all data (not limited to existing report types)
→ Generative UI renders visualisation widgets in real time as conversation progresses
→ Existing dashboards still accessible but visually deprioritised

**Drill Down & Control**
→ Clicking a visualisation opens a dedicated report with manual controls
→ Floating agent remains available for continued querying within that context

**Evolve & Act**
→ Agent proactively monitors KPIs, flags anomalies, and surfaces recommendations
→ Connected to task system — insights generate actionable next steps
→ Gateway to full home page hub-and-spoke integration

---

### The Merge Strategy

- **Approach A ships first** → faster path to closed beta, real usage data collected
- **Approach B built in parallel** → longer-term conversational-first vision
- Approach A feeds learnings into Approach B and ultimately unlocks the home page
- **End state (end of 2026):** Home page orchestrates all agents; data questions route to the insights agent, preserving manual controls alongside natural language