Design an AI Meeting Scheduler Agent

Act as a productivity systems architect. Design an AI meeting scheduling agent for [COMPANY NAME] with [NUMBER] team members.

**Context:**
- Calendar system: [GOOGLE CALENDAR/OUTLOOK]
- Team size: [NUMBER] people with [NUMBER] average meetings per week each
- Meeting types: 1-on-1s, team meetings, client calls, [OTHER]
- Working hours: [START] to [END] in [TIMEZONE]
- Buffer time required: [MINUTES] minutes between meetings

**Deliverables (numbered):**
1. Agent architecture: calendar integration, availability computation, preference learning, negotiation protocol, and booking execution
2. Availability algorithm: find mutual free slots considering working hours, timezone differences, existing meetings, buffer requirements, and focus time blocks
3. Preference learning: track which times each person prefers (morning vs. afternoon), which meetings they consistently decline, and optimal meeting length by type
4. Negotiation flow: when external participants are involved, send smart scheduling links with top 3 options, handle counter-proposals, and auto-confirm when all accept
5. Rescheduling intelligence: detect conflicts (meeting overlaps, participant OOO), proactively suggest alternatives, notify all parties with reason
6. Meeting preparation: 24 hours before, send agenda reminder, pull relevant documents from [SOURCE], and list action items from previous meeting
7. Post-meeting: detect no-shows, reschedule automatically, log meeting duration vs. scheduled, and suggest optimal duration for recurring meetings

**Constraints:**
- Must never schedule during marked OOO or focus time
- Must respect 'no meetings on [DAY]' policies
- Must handle recurring meeting pattern changes gracefully