Identity Stack

Your AI receptionist is customized across three layers, each editable independently from your dashboard. Changes take effect on the next inbound call.

The Three Layers

These three combine with three automatically-built layers (live Square data, session context like current time and status, client memories from previous calls) to produce the prompt the AI uses on each call.

Personality

The personality layer is short — a few paragraphs — and sets the how of every sentence the AI speaks.

Example (high-end salon):

You are Marissa, the receptionist at Luxe Midtown. You speak with
warmth and polish. You never rush the caller. You remember that this
salon's clients expect to be treated like the regulars they are, even
on their first call.

Example (neighborhood barbershop):

You're the front desk at Mike's. Keep it short, keep it friendly, no
unnecessary small talk. Match the caller's energy. Always confirm the
barber before booking.

The agent's name is set here. So is its language style, its tone with different caller types (rushed, chatty, nervous, frustrated), and any taglines or phrases you want it to use consistently.

Procedures

The procedures layer defines call flows — the numbered steps the AI follows for each situation.

FrontdeQ ships with a default procedures layer covering:

• New Booking — location, service, stylist, availability, contact details, confirmation
• Reschedule — identify the booking, find new availability, update
• Cancel — identify the caller, list bookings, confirm, cancel (see Cancel Flow)
• Pricing Questions — always use list_services, never guess
• Walk-ins — search today immediately, offer alternatives if full
• Escalation — when to transfer to a human, what to do when transfer fails (take a message)
• Voicemails — when salon is closed, what to collect, urgency levels
• Interruption handling — barge-in, mid-booking context changes, re-opened topics

You don't rewrite these from scratch. You add salon-specific overrides:

## Walk-ins
Only take walk-ins with Sofia on Thursdays. For any other day, tell
the caller we don't accept walk-ins and offer a same-day booking if
available.

Overrides are additive — they append to the default procedures. If you need to change a default step, edit that specific step.

Salon Knowledge

The knowledge layer is the most concrete — the facts your AI needs to answer questions that aren't a simple booking.

Organize it however is easiest to read. A common structure:

• Location & Parking — address, parking instructions, nearest landmarks
• Hours — regular hours, holiday schedule, seasonal variations
• Staff — each stylist, their specialties, schedule patterns
• Policies — cancellation, late arrival, refunds, gift cards
• Services beyond Square — anything bookable outside Square (weddings, events, consultations)
• Products — brands carried, bond treatments, color lines
• FAQs — the questions every caller asks (do you take walk-ins? do you have parking? is it kid-friendly?)

The AI pulls from this layer whenever a caller asks a factual question. If the knowledge isn't here, the AI will say "That's a great question — let me have someone get back to you on that" and take a message.

Editing the Layers

From your dashboard:

1. Settings → Integrations → Identity Stack
2. Three tabs: Personality, Procedures, Salon Knowledge
3. Each tab is a rich text editor with version history in the sidebar
4. Click Save — the change is deployed to your live agent within seconds

You can edit one layer without touching the others. You can roll back a single layer to any prior version without affecting the other two.

Versioning

Every save creates a new version. The sidebar shows:

Personality layer
  v4 — "Softened greeting for first-time callers" — 3 days ago — you
  v3 — "Added warmth with frustrated callers" — 1 week ago — system
                                                             (nightly QA)
  v2 — "Initial customization" — 2 weeks ago — you
  v1 — "Default" — 3 weeks ago — system

Each version shows:

• Version number
• Commit-style description (you write it when saving; system-generated suggestions include their rationale)
• Timestamp
• Author (you, a specific team member, or the nightly QA system if you accepted an auto-proposed fix)

Click any version to see its full contents. Click Rollback to make it live.

Integration With the Self-Improvement Loop

The nightly QA system scores every call and proposes specific edits to these layers. Suggestions include:

• Which layer the fix targets
• The exact text to add
• How many calls would have been handled better
• Why the system thinks this change helps

You approve suggestions in Call Quality → Suggestions. Accepting one applies it to the relevant layer, creates a new version, and deploys the change. See Owner Notifications for how accepted suggestions show up in the nightly recap email.

Starting Point

The default layers ship good enough to take calls on day one. You don't need to customize anything to go live. Add salon-specific tuning as you hear the AI on real calls:

1. Knowledge first. Add parking, policies, staff specialties, and FAQ answers as you notice what's missing. Highest-impact, easiest to write.
2. Personality second. Tune the tone once you've heard the AI on a few dozen calls. You'll notice it's too formal or too casual — edit accordingly.
3. Procedures last. Leave the defaults alone unless you have a specific flow you want to change. Editing without a reason usually makes things worse.

Let the QA loop surface what's missing. The system is good at finding patterns you wouldn't notice on your own.