Phase 4 priority shift: Meg needs AI assistant NOW (she is staff member #1)

CRITICAL INSIGHT:
- Meg = The Emissary = Current staff member #1 (not future)
- Meg doesn't have Claude access
- Every Meg question = Michael interruption
- AI assistant = Meg self-serves 24/7

TIMELINE CHANGE:
- Was: "Deploy when staff wiki exists" (Month 4+)
- Now: Deploy immediately after Phase 3 (Session 6)
- Impact: Meg gets AI assistant 4-6 months earlier

PHASE 4 UPDATED:
- Renamed: "Staff AI Assistant for Meg" (not generic staff)
- Knowledge Base: Emissary-focused (~20-30 docs)
  - Social Media Handbook
  - Consultant Profiles
  - Subscription Tiers
  - Contact Reference
  - Origin Story
- Questions Meg can ask:
  - "What personality traits for Jack in social posts?"
  - "When to post Fire vs Frost content?"
  - "How to describe Awakened tier?"
  - "Who handles billing issues?"
- Training: 15 min ("type question, get answer")

BENEFITS:
- Reduces Michael interruptions immediately
- Builds Meg's tech confidence (success → NextCloud later)
- Proves concept before recruiting more staff
- Recruitment advantage: "We have AI assistant"
- Simple interface (accessibility win)

SESSION 6 UPDATED:
- Added: Deploy Meg's AI assistant (2 hours)
- Added: Train Meg on usage (15 min)
- Total time: 4-5 hours (was 2-3)

Updated by: Chronicler the Ninth

docs/core/tasks.md

@@ -401,30 +401,80 @@ git clone https://git.firefrostgaming.com/firefrost-gaming/brainstorming.git
 
 ---
 
-**Phase 4: Staff AI Assistant (2-3 hours)**
+**Phase 4: Staff AI Assistant for Meg (2-3 hours)**
 
-**Deploy Open WebUI with staff wiki docs only**
-- Much smaller dataset (~50-100 docs when staff wiki exists)
-- Built-in Chroma vector DB sufficient (no need for external)
-- Embedded chat widget OR dedicated portal
-- Domain: staff-ai.firefrostgaming.com
+**PRIORITY SHIFT:** Meg is staff member #1. Deploy immediately after Phase 3 (not "when staff wiki exists")
+
+**Why Meg Needs This Now:**
+- Meg = The Emissary = Staff member #1 (not future staff, current staff)
+- Meg doesn't have Claude access
+- Meg needs to self-serve answers about:
+  - Social media posting guidelines (Fire vs Frost content)
+  - Consultant personality traits for posts
+  - Subscription tier descriptions
+  - Who to contact for what (Michael? Breezehost? Payment processor?)
+  - Basic procedures and workflows
+- **Every Meg question = Michael interruption**
+- AI assistant = Meg self-serves 24/7
+
+**Deploy Open WebUI for "Emissary Knowledge Base"**
+- Start small: ~20-30 docs (not 50-100, grow over time)
+- Built-in Chroma vector DB sufficient (small dataset)
+- Simple web interface (easier on Meg's tech comfort)
+- Domain: staff-ai.firefrostgaming.com (or emissary-ai.firefrostgaming.com)
+
+**Initial Knowledge Base (Emissary Focus):**
+1. **Emissary Social Media Handbook** (already exists in `docs/planning/`)
+   - Fire path vs Frost path content strategy
+   - Posting schedule, content pillars
+   - Platform-specific guidelines
+2. **Consultant Profiles** (`docs/relationship/consultant-profiles.md`)
+   - Jack, Oscar, Butter, Jasmine, Noir personalities
+   - Photo reference for social media posts
+   - Lore and character traits
+3. **Subscription Tiers** (`docs/planning/subscription-tiers.md`)
+   - Tier names, prices, benefits
+   - How to describe each tier in marketing
+4. **Basic Contact Reference**
+   - Who handles what (Michael = infrastructure, Breezehost = hosting, etc.)
+   - Emergency contacts
+   - Escalation paths
+5. **Origin Story** (`docs/relationship/origin-story.md`)
+   - Brand storytelling reference
+   - How Michael & Meg met (Donna's Restaurant)
 
 **Configuration:**
-1. Create "Staff Wiki" knowledge base in Open WebUI
-2. Upload staff-facing docs only (operations manual stays private in AnythingLLM)
-3. Configure access (staff accounts, not public)
-4. Test 24/7 staff question answering:
-   - "How do I restart a game server?"
-   - "What's the whitelist process?"
-   - "Who do I contact for billing issues?"
-5. Document usage in staff wiki
-6. Train Meg on basic usage
+1. Create "Emissary Knowledge Base" in Open WebUI
+2. Upload Meg-relevant docs only (operations manual stays private in AnythingLLM)
+3. Set up Meg's account (simple username/password, no complicated auth)
+4. Test with real Meg questions:
+   - "What personality traits should I emphasize for Jack in social posts?"
+   - "When should I post Fire path content vs Frost path content?"
+   - "How do I describe the Awakened tier to potential subscribers?"
+   - "Who do I contact if someone has a billing issue?"
+   - "What's the origin story of Firefrost Gaming?"
+5. Train Meg on usage (15 min: "Type question, hit enter, get answer")
+6. Monitor first week, add docs based on Meg's actual questions
 
 **Benefits:**
-- Reduces Michael/Meg interruptions (staff self-serve)
-- 24/7 availability (AI doesn't sleep)
-- Onboarding tool for future recruitment
-- Consistent answers (no "telephone game")
+- ✅ Reduces Michael interruptions immediately (Meg self-serves)
+- ✅ 24/7 availability (Meg can ask at 2 AM if she's working)
+- ✅ Builds Meg's confidence with tech (success with AI → confidence for NextCloud later)
+- ✅ Proves concept before recruiting more staff
+- ✅ **Recruitment advantage:** "We have AI assistant" = professional operation
+- ✅ Consistent answers (AI doesn't forget, doesn't give conflicting info)
+- ✅ Foundation grows: Add docs as staff grows, knowledge base scales naturally
+
+**Accessibility Win:**
+- Simple interface (type question, get answer)
+- No complex menus or navigation
+- Mobile-friendly (Meg can use on phone)
+- No Git, no terminal, no technical barriers
+
+**Timeline:**
+- **Was:** "When staff wiki deployed" (Month 4+)
+- **Now:** Session 6, immediately after AnythingLLM ingestion complete
+- **Impact:** Meg gets AI assistant 4-6 months earlier
 
 ---
 
@@ -1176,12 +1226,14 @@ TIER 6: FUTURE (Month 4+ or 2027)
 2. Start model downloads (overnight)
 3. **Checkpoint:** Downloads running, come back tomorrow
 
-### Session 6 (2-3 hours): AI Stack Complete
-1. Verify models loaded
-2. Gitea integration (1 hour)
-3. Staff AI assistant setup (2 hours)
-4. Test DERP functionality
-5. **Checkpoint:** Full AI stack operational
+### Session 6 (4-5 hours): AI Stack Complete + Meg's Assistant
+1. Verify AnythingLLM workspaces fully ingested
+2. Test DERP functionality (reconstruct session from repo)
+3. Deploy Open WebUI for Meg (2 hours)
+4. Create "Emissary Knowledge Base" (Social Media Handbook, Consultant Profiles, etc.)
+5. Train Meg on usage (15 min: show her how to ask questions)
+6. Test with real Meg questions
+7. **Checkpoint:** Full AI stack operational, Meg can self-serve 24/7
 
 ### Session 7 (1-2 hours): Monitoring
 1. Deploy Netdata (1-2 hours)