Aeion Concierge for VIP Customer Orchestration
When a high-value customer transacts, complains, or signs up, the response in the next 60 seconds determines whether they feel seen or invisible. Concierge's event-driven VIP Welcome + Sentiment Guard + B2B Lead Enrichment agents react in single-digit milliseconds — tagging the record, escalating to the right human, drafting a personalized response — without your team having to spot the moment. Every action lands in the Concierge audit log with full traceability.
The VIP Recognition Problem
Every business has a long tail of customers with wildly varying lifetime value. The top 5-10% generate the majority of revenue. Recognizing and responding to them in real time is the difference between retention and churn:
Walkthrough 1 — VIP Order Welcome
`
Walkthrough 2 — Angry Customer Detection
`
Walkthrough 3 — B2B Lead Enrichment
`
The Cross-Module Orchestration Pattern
`
Configuration Patterns
`
FAQ
Intercom's bot is great inside the Intercom chat surface. Concierge operates across CRM + Commerce + Inbox + Notifications simultaneously. Same conversation, more context, more action.
Single-digit milliseconds for the event-driven agents (VIP Welcome, Sentiment Guard, B2B Enrichment, Loyalty Liaison). Operational agents (Chief of Staff, Assistant, Travel) run on scheduled/on-demand triggers with seconds-to-minutes latency.
The 7 built-in agents cover the most common VIP-recognition and executive-productivity workflows. For automations beyond those, use Blueprint — our visual, no-code workflow builder — or talk to our solutions team about a bespoke agent as part of an Enterprise engagement.
The escalation trigger (negative sentiment OR urgency >= 4) is a fixed default today that errs on the side of escalation. Per-tenant tuning is intended for the agent's config record but isn't wired into the runtime logic yet — treat it as a shipped default rather than a dial you can turn from the admin UI today.
Sentiment Guard falls back to a keyword-based heuristic (angry/refund/cancel/terrible/furious/disappointed, etc.) instead of skipping analysis — the fallback is noted in the record so you can tell an AI call didn't happen, but the action still completes rather than failing. VIP Welcome and Lead Enrichment don't call AI at all, so they're unaffected either way. The three operational-input agents (Chief of Staff, Assistant, Travel) use the platform's strict honest no-op contract: no AI provider configured records a plain "no AI provider connected" reason and the action doesn't complete.
Concierge agents are channel-agnostic. The Sentiment Guard reacts to a new message regardless of source channel. The customer record aggregates all interactions across channels.
Customer messages go through your configured AI provider. For PII-sensitive industries, use the AI provider gateway's data-classification rules. Some tenants use on-prem LLMs (via gateway).
Yes. Concierge agents are optional — Blueprint lets you build deterministic, rule-based automations (no AI) using the same event triggers and audit trail. Use AI only where its strengths matter.
The Concierge audit log inherits the tenant's general audit log retention policy: 30 days on Starter, 1 year on the Professional baseline (what most tenants run), up to 7 years on Enterprise.
Concierge itself is included free with any Aeion module — no separate subscription. You only pay for AI inference (per-agent, usage-based). Most tenants spend $100-500/month on AI inference depending on event volume.