Onyx
GLAMOUR AI
ONYX

Sign in to your AI Workspace

Manage AI leads, conversations, missions and results in your organization workspace.

No organization API key is used in this browser.
Onyx
Glamour AIAI Workspace
Glamour AIAI Workspace
TENANT OPERATIONS

Overview

DEMO ENVIRONMENTSynthetic sales data only · external business delivery and real phone-network calls are blocked.
Reset from server CLI
ONYX LIGHT · DAILY WORKFLOW

Good morning.

Here is what needs your attention today.

O
LEAD JOURNEY

See where every opportunity is moving.

Each stage is clickable. Light keeps the workflow visible without exposing the system underneath.

0 active
NEEDS ATTENTION

What to do next

RECENT ACTIVITY

What just happened

LEADS

Every lead, without the clutter.

Search or choose a journey stage. Open a lead to see one clear story and the next best action.

LEAD STORY

Lead

AI SCORE0/100
WHAT ONYX KNOWS

Key facts

JOURNEY

Everything in one timeline

ONYX AI CALL

Talk to a lead with Onyx.

Choose the lead. Onyx automatically uses the saved language and lead context. Everything else stays in the background.

Checking voice…
QUICK CALL WORKFLOW

Tell Onyx what this call should achieve.

Optional. Pick a shortcut or describe the goal in plain language. Onyx keeps platform safety, verified facts and the selected lead context in control.

Standard call
CHOOSE A LEAD

Who should Onyx speak with?

0 leads
RECENT CALLS

Latest AI conversations

SELECTED LEAD

Lead

LIVE CONVERSATION

Customer ↔ Onyx

The conversation appears here turn by turn while the AI call is running.

Live transcript

Start the call to see the conversation here.

ACTION CENTER

Only the work that needs a human.

Follow-ups, viewing confirmations and other open actions stay in one short queue.

0 open
ACTIVITY

A readable history of the workspace.

Recent AI and team events are translated into a simple feed. Full audit controls remain available in Pro.

SETTINGS

Keep daily settings simple.

Your profile and workspace context live here. Advanced organization, AI and security controls remain in Pro for authorized roles.

YOUR WORKSPACE

User

WORKSPACE MODE

Light for focus. Pro for full control.

Lead pipeline

Recent activity

LIGHTWEIGHT CUSTOMER CONTEXT

AI Workspace

Understand each lead across identity, requirements, AI work, Website Chat, linked Voice activity and internal notes without turning Glamour AI into a full CRM.

Export CSV

Add contact and lead

Leads

Open a lead to see its complete AI context.

0 leads
LEAD WORKSPACE

Lead

AI lead score0/100
BROWSER AI CALL

AI Call

Uses the lead's saved language and context. Browser microphone/WebRTC is used only when a real realtime provider is configured; phone-network calling is not part of this stage.

Ready
Call time00:00
Select Start AI Call to issue a short-lived, origin-bound browser ticket. No organization API key is exposed to this browser.

Live call events

No active call.

Live transcript

Transcript events will appear here when the realtime provider sends them.

Post-call intelligence

After a call ends, Glamour AI analyzes finalized transcript data into canonical English business context. No semantic facts are invented when a real text provider or transcript is unavailable.

Not analyzed

End a linked call to run the automatic post-call pipeline.

Contact context

Requirements

Internal notes

Workspace-only context for your AI team.

Activity timeline

Website conversations

Only sessions matched inside this organization by verified captured phone/email identity are shown.

Linked AI voice calls

Only voice sessions explicitly linked to this lead are shown.

Viewings & appointments

Requested does not mean confirmed. Confirmation source and external sync status remain visible.

Seller & landlord acquisition

Evidence-backed acquisition cases explicitly linked to this lead. Relationship, outreach rights and listing status are never assumed.

AI work & results

Missions and results already associated with this lead.

BROWSER VOICE

Voice AI

Start lead-aware AI conversations inside the browser. Glamour AI keeps server credentials off the device, binds each session to the organization and lead, and records the Voice session in the AI Workspace.

Loading…

Browser voice profiles

Realtime readiness only. Profile editing remains controlled by the existing Voice configuration APIs.

What this stage does

TransportBrowser microphone + WebRTC when configured
Lead contextTenant-isolated and linked explicitly
LanguageUses the lead's saved conversation language
Post-call intelligenceTranscript → summary → score → requirements → next action → workspace/CRM
Phone networkNot enabled — Telephony remains separate
Browser secretsNone — short-lived ticket only

Recent browser AI calls

Only sessions explicitly linked to a lead in this organization are shown.

0 calls
PHONE NETWORK

Telephony

Prepare real inbound and outbound AI phone calls without exposing provider credentials. Browser AI Call stays separate and continues to work even when no phone-network provider is configured.

Checking provider…

Provider readiness

A real phone call is enabled only when an external-network provider, active connection, dedicated number and AI route are all ready.

Registered providers

Simulation providers never count as production phone-network readiness.

No fake phone calls: if this environment only has mock_telephony, Glamour AI will show Provider setup required and will not enable the real outbound call button. Provider credentials remain server-side and are referenced, never pasted into browser code.

Provider connection

Dedicated number

AI route

Current phone configuration

Public connection data never includes raw provider credentials.

Call a lead

Inbound phone policy

Inbound activationProvider webhook / adapter required
AI routingDedicated number → tenant route → AI profile
LanguageRoute language or AI profile default
CredentialsServer-side only

Recent phone-network calls

Real provider calls will appear here when Telephony is configured. Simulation records remain clearly labeled by provider.

0 calls
INTERNAL SALES INTELLIGENCE

Agent Copilot

Ask about one lead using that lead’s real Workspace context. Copilot can summarize history, recommend the next best action, draft a customer message, and suggest only verified published properties. It does not send messages, place calls, confirm viewings or modify CRM records.

Loading…

Lead context

Ask Copilot

Verified property candidates

Only Verified + Published catalog items can be recommended. Empty means Copilot must not invent a property.

0

Copilot conversation

History is scoped to this signed-in user, organization and selected lead.

0 messages
CORE WEBSITE AI

Website AI Agent

Configure the AI agent that lives on your website, qualifies visitors, captures leads and keeps each conversation inside this organization. CRM is optional and no server secret is exposed to the browser.

Loading…

Agent & website connection

Connection & security

Public widget identifiers are safe to embed; organization API keys remain server-side.

Widget preview

Visual preview only. It does not create a fake AI conversation.

Website install

Use the public channel key with the existing Glamour Chat SDK. Never place a server API key in website code.


                

Recent website conversations

Real sessions from this organization only. Open one to inspect messages, qualification, requirements, lead capture and usage.

0 sessions
WEBSITE CONVERSATION

Conversation

Messages

AI context

Lead capture

VERIFIED PROPERTY DATA

Property Knowledge

Give Glamour AI a safe property catalog even when Glamour Map is not connected. Structured property facts are stored in English; AI conversations can still happen in the customer's language.

Loading…

Active data sources

Source strategy is managed under AI Configuration → Data Sources. Property Knowledge stays focused on verified content and grounding.

Add or update property

CSV paste import

Paste up to 500 rows. Required headers: external_id,title. Common optional headers include property_type, purpose, city, community, project, developer, bedrooms, bathrooms, price, currency, area_sqft, availability_status, summary and target_url.


            

Property catalog

Only verified + published records from trusted active sources are eligible for AI grounding. Availability must remain explicit.

0 properties

Grounding search test

Test the same verified-only knowledge search used by Glamour AI. This does not call a text AI provider.

Paste raw customer numbers

One row per customer: phone | name | language | status | notes


          

Create reusable mission

Missions

Create work item

This mission is already queued or running for the selected lead. Cancel the pending item first or choose another mission.

Operational queue

Completed AI results

Add workspace user

Create a secure dashboard account using one of the Glamour AI access levels. Owner is the protected highest level.

Access levels

Organization team

Customer-facing roles stay intentionally simple: Owner, Admin, Manager and Agent. Owner accounts are protected from Admin changes.

SAFE SCHEDULING AGENT

Viewings & Appointments

Capture viewing requests, confirm them explicitly, reschedule safely, create reminders and keep post-viewing follow-up tied to the same lead. A requested time is never presented as confirmed.

Loading…

Create appointment request

Confirmation policy

RequestedCustomer preference only — not confirmed
ConfirmedExplicit human verification or real external calendar event
RescheduleInvalidates the previous confirmation
ReminderInternal Follow-up Engine action before a confirmed appointment
Post-viewingSuggested follow-up after completed/no-show viewing
External syncFail-closed when no active integration exists
Safe action claim: Glamour AI may say a viewing was requested. It may say confirmed only after the scheduling record has an explicit confirmation source.

Appointment queue

All status changes remain in appointment history. Confirmed items can create reminders; completed viewings can create post-viewing follow-up suggestions.

0 appointments
ENGAGEMENT AI

Follow-up Engine

Keep AI-recommended and team-created next actions clear, timed and auditable. This stage schedules work; it does not pretend an external message or phone call was sent.

Checking Text AI…

Schedule follow-up

Execution policy

Human-approved by design: Browser Voice requires a user gesture and microphone permission. WhatsApp, email and real phone delivery are not enabled by this stage and will only appear when their real integrations are configured.

Follow-up queue

AI recommendations from completed Post-call Intelligence arrive as Suggested. A team member approves them before they become Scheduled.

VERIFIED PROPERTY MARKETING

AI Listing Studio

Generate multilingual listing and marketing copy from verified property facts only. Missing facts are omitted, never invented, and every generated version remains auditable against the property checksum used at generation time.

Loading…

Generate content bundle

OutputsListing title · Portal description · Short copy · Luxury copy · Social caption · Key features · CTA
A real server-side Text AI provider and at least one verified property are required. No API key is exposed to this browser.

Fact safety policy

SourceVerified Property Knowledge only
Missing factOmit it — never infer it
Price / availabilityUse only when present in verified facts
Views / amenitiesNo invented marketing claims
ROI / appreciationNo guaranteed returns
ApprovalBlocked if property facts changed after generation
Approval is not publishing: approving a version marks the copy as internally approved. This stage does not publish to portals, social networks or third-party systems.

Content preview

Select a generated version to review every output and its verified-fact audit state.

No version selected

No generated listing version selected.

Version history

Every generation is immutable. Approving a new version archives the previous approved version for the same property and language.

0 versions
GROWTH AI · AUTHORIZED DISCOVERY

AI Prospecting & Owner Discovery

Find and review real-estate opportunities from explicitly public or authorized sources, preserve source evidence, detect duplicates across sources, and promote reviewed opportunities into the AI Workspace without inventing private data or owner relationships.

Loading…

Authorized source

Discovery safety policy

SourcesPublic or explicitly authorized only
Domain scopeAI web search restricted to approved domains
Owner classificationRequires explicit source evidence
Contact dataCopy public/authorized details only — never infer private data
Access controlsNo login, paywall or CAPTCHA bypass
PromotionHuman action required before entering AI Workspace
Evidence first: a high prospect score means the opportunity is worth reviewing. It does not prove ownership, authority to market, availability or permission to contact. Those facts must remain evidenced and reviewed.

Run domain-restricted discovery

Real server-side web-search AI must be enabled. No API key or raw provider credential is exposed to the browser.

Add authorized observation

Prospect review queue

Cross-source duplicates are merged by deterministic fingerprint while every sighting remains auditable. Relationship labels stay Unknown unless directly evidenced.

0 prospects

Recent discovery runs

Each run records its source, criteria, provider state and counts for observed, new and duplicate opportunities.

Configured sources

Source authorization context remains visible for audit. Raw provider credentials are never stored here or sent to the browser.

GROWTH AI · LISTING ACQUISITION

Seller & Landlord Acquisition

Qualify owner or authorized-representative opportunities into real listing conversations. Keep ownership evidence, outreach rights, asking price/rent, availability, agency cooperation, exclusivity and viewing readiness separate from assumptions.

Loading…

Start acquisition case

Acquisition safety policy

OwnershipNever inferred without relationship evidence
OutreachHuman confirmation required before contact-ready
Availability / priceUnknown stays unknown until confirmed
Agency cooperationRecorded separately from ownership
Exclusive listingInterest is not a signed mandate
External messagingNo automatic WhatsApp/email/phone delivery in this stage
Commercial truth: a qualified acquisition case means the opportunity is worth progressing. It does not mean the company has an exclusive listing agreement, permission to market, or confirmed viewing access unless those facts are separately verified.

Qualification facts

Approved next contact

Only Internal Action and Browser AI Call are actionable here. Browser Voice still requires a user gesture; external phone/WhatsApp/email must use their real configured integrations.

Acquisition pipeline

Qualification score is explainable. Human status transitions remain explicit, and Do Not Contact is terminal.

0 cases
DORMANT LEAD RECOVERY

Lead Reactivation

Find genuinely dormant leads from real interaction history, review them before outreach, and return a lead to the active sales team only after verified renewed interest. Old requirements are background context and must be reconfirmed.

Loading…

Create dormant-lead scan

Dormant statuses
No automatic outreach. Creating a campaign only scans tenant-isolated lead history. Every reactivation action still needs human approval.

Recovery safety

Do Not ContactFail-closed
Old requirementsMust be reconfirmed
Inbound replySignal, not proof of interest
Hot recoveryVerified renewed interest only

Notes or ordinary record edits do not reset inactivity. Last engagement is derived from linked Website Chat, Voice, completed Viewings and verified acquisition contact history.

Approve reactivation action

Approval creates a Follow-up Engine action. It does not send an email, WhatsApp message or phone call automatically.

Record verified response

Only Renewed interest returns the lead to Qualifying + High priority. A new inbound message is detected as a response signal but still requires review.

Reactivation candidates

Response detected is not automatically Hot. Review the new evidence before recording renewed interest.

0 candidates

Campaign history

Campaign scans are preserved for audit and do not delete old lead history.

0 campaigns
PERFORMANCE & AI QUALITY

Analytics

Measure what Glamour AI can prove from this organization's stored records: lead activity, channels, follow-ups, viewings, growth workflows, usage/cost estimates and reviewable AI quality signals.

No synthetic business claims: this page does not invent revenue, ROI, closed-deal conversion or a made-up AI quality score. Those metrics are shown only when a real system of record can support them.

Lead pipeline

Lead records touched inside the selected reporting period.

Lead sources

Where newly created workspace leads came from.

Website AI

Voice & post-call

Follow-up & viewings

Growth AI

AI usage & estimated cost

Estimated cost reflects recorded provider usage events only; it is not an invoice.

AI Quality Review

Turn stored guardrail and failure evidence into a review queue. Quality flags are auditable signals, not a synthetic score.

WHITE-LABEL WORKSPACE

Organization

Organization profile

White-label branding

Workspace configuration

Organization identity, commercial plan and workspace ownership stay together. AI behavior is configured separately under AI Configuration.

AI CONFIGURATION

AI Identity

Define how this organization's AI presents itself. Customer-facing conversation language can change per lead; canonical structured data remains in English.

Organization-scoped

Identity & behavior

AI SETTINGS

Data Sources

Choose where Glamour AI gets verified property information. Turn on only the sources this organization actually uses; each source reveals only its own settings.

No browser secrets
Architecture rule: Glamour AI never shares a property database with Glamour Map or another product. Website/API and Map stay behind integration boundaries; AI-hosted data remains tenant-isolated inside this AI service.
ORGANIZATION PACKAGE

Plan & Features

One place to understand what this organization purchased and which Glamour AI modules are available, enabled or locked.

Loading…

Feature entitlements

Glamour assigns the purchased plan and platform availability. The Organization Owner may toggle only optional modules that are purchased and unlocked.

Single commercial view: Plan/Tier metadata and live module entitlements are shown together here. Organization Admins cannot change the purchased plan; Glamour controls plan assignment and platform availability.
AI CONFIGURATION

Prompt Architecture

Glamour AI uses ordered prompt layers for safety, personality, greeting, qualification, property advice, follow-up, handoff, closing and structured results. Customer organizations can review the architecture but cannot alter protected platform behavior.

Protected prompts

Prompt templates & versions

Published platform templates are the controlled source of behavior. Organization-specific legacy templates remain visible for audit but are protected from customer editing.

GLAMOUR PLATFORM CONTROL

Edit protected prompt

Only the authorized Glamour Platform Owner can create a new draft and publish it. Customer Organization Owners and other Glamour users cannot use this editor.

Version safety: saving never overwrites the currently published version. A new draft is created first; publishing is a separate action so the active prompt always has an auditable version history.

Version history

Governance: platform safety and core conversation behavior are controlled by Glamour. Company-specific identity, AI name, language and approved tone variables are managed under AI Identity instead of rewriting protected system prompts. No API key is sent to the browser.
SAFE REAL ESTATE ADVISOR

Safety & Advice

Mandatory real-estate guardrails keep verified facts separate from analysis, prevent unsupported guarantees, and route sensitive legal, tax, residency and financing topics toward cautious information and human follow-up.

Loading policy…

Safety classifier test

Test how Glamour AI classifies a customer question before a real provider answers. This is deterministic policy inspection, not a fake AI conversation.

Policy decision

Run a question to see detected topics, safe next action and grounding requirements.

Mandatory policy: customer organizations cannot disable these safety controls. The customer may speak in any supported language; original conversation text stays intact for audit, while derived operational data remains canonical English.
INTEGRATION CONTROL PLANE

Integrations Center

Manage CRM, Glamour Map, Website/API and signed webhook boundaries from one tenant-isolated control center. Raw credentials are never stored or shown in the browser.

Loading…

Integration catalog

A status overview of supported connection families. “Ready” is configuration readiness, not a fabricated network-health claim.

API boundary rule: Glamour AI, CRM, Map and Properties never share a database. Connections use canonical APIs/webhooks, tenant isolation, scopes, signatures and idempotency.

Configure connection

Connections

Configuration readiness is deterministic and does not make an external network request.

Readiness history

Signed outbound webhook

Webhook endpoints

Fail-closed delivery: external webhook delivery only leaves the server when the production delivery flag is enabled. Otherwise queued deliveries remain safely queued.

Recent sync jobs

Webhook delivery queue

GLAMOUR AI GUIDE

Workspace Help & Guide

A simple living guide for the dashboard. As new Glamour AI modules are added, their instructions will be added here too.

Always available

Quick start

  1. Add or receive a lead.Use AI Workspace, Website AI Agent, imports or an integration.
  2. Choose the AI job.Create or select a Mission such as qualification, requirement extraction or next-action recommendation.
  3. Run and review.Work Queue holds operational tasks; Results stores completed AI outcomes.
  4. Keep the workspace controlled.Owners/Admins manage organization settings; Owners protect organization ownership.

Dashboard sections

Select any section below for a short explanation.

Access levels

Owner

Highest authority. Can manage the full workspace and create/assign another Owner. Admins cannot change, deactivate or replace an Owner.

Admin

Runs organization settings, team access, integrations and AI operations, but cannot manage protected Owner accounts.

Manager

Manages day-to-day operations, missions, imports, queue and results with team visibility but without sensitive organization changes.

Agent

Works with leads and operational AI tasks without organization or team administration.

Safety rule: the last active Owner cannot be deactivated or demoted. Assign another Owner first if ownership needs to change.