Aeion Concierge for Executive Briefing Automation
Executives lose 5-10 hours a week to context-gathering: triaging overnight email, pulling meeting context, juggling travel logistics. Concierge's three operational agents (Chief of Staff, Assistant, Travel) turn structured input into a synthesized briefing, dossier, or logistics read — wire each one to a Blueprint trigger (a morning cron, a pre-meeting schedule, an itinerary check) and the result lands in the Concierge audit log every time. Confidence-flagged so high-stakes calls stay human in the loop.
The Executive Time-Drain Problem
`
Walkthrough 1 — Chief of Staff Morning Briefing
`
Walkthrough 2 — Assistant: Meeting Prep
`
Walkthrough 3 — Travel: Logistics Monitoring
`
Confidence Flagging — Human In The Loop
`
Three Operational Patterns
`
FAQ
The email content your workflow feeds into Chief of Staff — subject lines and a preview (not the full body) — goes to the AI to generate the briefing. Use the AI provider gateway's data-classification rules for sensitivity. Per-tenant data residency configurable.
Travel Agent is advisory only today — it outputs a recommendation and a suggested buffer, it never touches your calendar directly. If you want an automatic calendar update, chain that as a separate step in the Blueprint workflow that calls Travel Agent, gated however you like.
There's no built-in per-user scheduler today — you set the cadence via the Blueprint cron trigger that calls the briefing action, so timezone, hour, and weekend handling are whatever you configure on that trigger.
Assistant Agent itself is calendar-agnostic — it takes a meeting title, time, agenda, and attendee list as structured input, however your workflow sources it. The platform's own calendar sync (used elsewhere, e.g. Appointments) supports Google Calendar and Microsoft Outlook today.
The itinerary schema categorizes legs as flight / train / car / rideshare / transit / other — Travel Agent doesn't call any airline, hotel, or rideshare API directly. Real-time transit and weather alerts are supplied by whatever workflow or integration you wire in front of it.
Attendee data passed to AI is limited to what your workflow supplies: name, org, role, and free-text CRM/support context. Per-tenant data-masking rules can restrict further.
Confidence thresholds are a fixed default (60) today across all three agents. Scheduling is whatever cadence you set per exec in your own Blueprint trigger.
Audit log shows the full prompt + response for every run. User overrides briefing items and can dismiss or promote flagged ones — both actions are recorded in the audit trail.
Concierge complements human EA. EA gets the briefing + dossiers too; EA decides what to escalate. Many tenants run with both.
Larger prompts (more emails, more attendees) increase cost. Per-tenant cost caps in AI provider gateway. Typical: $100-2000/month per exec.
Pause the agent's triggering Blueprint workflow, or delete/skip the run — there's no built-in PTO-aware pause today; that's on you (or your Blueprint automation) to gate.