8x8 Messaging Platform — Capability Report

Prepared for company leadership · June 2026

Executive summary

Blu Coffee Distributors runs a full 8x8 (CPaaS) messaging integration built directly into Odoo, giving us a single, governed platform for reaching customers and coordinating internal operations across Viber and SMS. The platform spans three capability areas: (a) Viber marketing — targeted, rich-media promotional blasts with SMS fallback — backed by a built-in AI ChatGPT chatbot that answers public inquiries 24/7 and automatically captures buying-intent prospects as CRM leads; (b) branded SMS sent under recognizable Blu sender identities (Orders, Service, Accounting, and more); and (c) a layer of automatic operational notifications that push real-time updates across Service, Purchasing, and Inventory workflows. Artificial intelligence is a first-class part of this platform: an on-brand ChatGPT concierge handles routine FAQs without staff involvement, and a second AI pass screens every inbound message for purchase intent and routes hot prospects straight into our sales pipeline.

Platform at a glance

Dimension	Detail
Provider	8x8 (CPaaS — Communications Platform as a Service)
Channels live	Viber and SMS (with automatic Viber→SMS fallback)
Branded sender names	Blu Coffee, Blu Orders, Blu Service, Blu Acctg, Blu Intrnal (routed via dedicated 8x8 sub-accounts)
AI	ChatGPT concierge (default model gpt-4o) for FAQ auto-reply + buying-intent → CRM lead capture
Delivery tracking	Per-recipient status (queued → sent → delivered → read, or failed) via 8x8 webhooks
Where it runs	Native inside Odoo (blu_8x8 module); managed by staff with no developer involvement

Marketing & customer engagement

Viber marketing campaigns

Marketing can run targeted, trackable Viber promotions entirely self-service — turning a curated customer list into a measured, rich-media blast in minutes.

• Build the right audience, fast. Recipient lists are assembled three ways: pick contacts individually, pull in everyone with a chosen tag (e.g. all "Wholesale" or "Cafe Owner" partners), or paste/upload a CSV of phone numbers with no contact record needed. Duplicates are blocked and numbers auto-formatted.
• Send rich content, not just text. A single blast can carry plain text, an image, a video, an audio clip, or a file (promo flyer, product reel, price list). If a recipient is not on Viber, the message automatically falls back to SMS using a separate, readable fallback message — so the audience is never left out.
• Test before you blast. A one-click "Send Test" fires the exact message to your own phone (or any number you type) to confirm wording, formatting, and image rendering before reaching real customers. Test sends do not touch or consume the actual recipient list.
• Immediate or scheduled delivery. "Send Now" sends right away, paced one recipient at a time to respect rate limits; "Put in Queue" hands large or overnight sends to a background process without tying up the user.
• Track delivery per person. Every recipient shows live status as it moves through queued → sent → delivered → read (or failed), updated automatically from Viber. Once everyone has received the message, the campaign marks itself Done — a clear, auditable view of reach.

AI chatbot & buying-intent capture

The Viber channel is more than a broadcast tool — it is a 24/7 AI-assisted sales front door.

• 24/7 AI concierge on Viber. Public inbound inquiries get an instant, on-brand ChatGPT reply with no staff in the loop, covering routine FAQs (products, services, pricing) the moment a prospect reaches out — no waiting for office hours or a free agent.
• A knowledge base our team controls. Answers are grounded in a curated FAQ table (Prompt → Answer pairs) that marketing edits directly in Odoo. Adding a row or correcting an answer takes effect on the very next inbound message — keeping the bot accurate and on-message without a developer or release cycle.
• Automated buying-intent capture ("Da Vinci Mode"). A second AI pass screens each inbound message for purchase intent. When it detects a hot prospect, it automatically creates a CRM lead — auto-linked to an existing customer record when the phone matches — so genuine opportunities are captured around the clock and never lost in an unread channel.
• Tap-to-claim lead racing. Each AI-detected lead is pushed to every member of the designated sales group via a Viber message with a one-tap "Yes, let me handle it" button. First to tap wins and owns the lead; the rest see an "already being handled" notice — driving fast follow-up and clear ownership with zero manual dispatching.
• Auditable and cost-aware. Every AI reply is logged as a record flagged is_ai_reply, providing a full audit trail of what the bot told customers. A response-length cap controls token cost, and the persona is selectable — all manageable from the settings page.

Branded SMS

• Branded sender identities on every SMS. The standard Odoo SMS composer is extended with a Sender Account selector, so each message is sent under a recognizable Blu identity — Blu Coffee (general marketing), Blu Orders (order/dispatch), Blu Service (repair/pickup notices), Blu Acctg (statements, payment confirmations), or Blu Intrnal (internal staff blasts) — all routed through 8x8.
• Single or bulk send, no new tooling to learn. Staff send from any contact's form via Actions → Send SMS Text Message, or in bulk by selecting multiple rows in the contacts list. It is the familiar Odoo workflow, with nothing extra to train on.
• Full audit trail with sender accountability. Every sent message is logged with its Sender Account shown as a column, so management can see exactly which team sent what, and a delivery-status webhook feeds back per-message states (sent/delivered/failed) for end-to-end traceability.
• Why branded senders matter. Recipients instantly recognize whether a message is an order confirmation, a payment reminder, or a promo — building trust and reducing the chance our texts are mistaken for spam or fraud. The per-team tagging also supports compliance, with clear sender attribution and an auditable log essential for handling consent, opt-outs, and any dispute over what was communicated.

Automated operational notifications

These are automatic, system-triggered messages — not marketing blasts. Each fires from a state change or a daily schedule, sends best-effort (failures are logged, never blocking the business flow), and de-duplicates recipient numbers before sending.

Workflow	Audience	Channel	Trigger	Business value
Service schedule / reschedule	Customer (portal user, their notification contacts, and partner mobile)	Viber ("Blu Service")	Repair order's schedule date or technician is set/changed while the service request is active	Customers know who is coming and when; cuts missed visits and "where's my technician?" calls
Post-service satisfaction survey	Customer	Viber ("Blu Service") with tokenized survey link	Repair marked complete — one survey per repair	Captures quality / hygiene / overall ratings tied to the specific repair and technician; closes the service-quality loop
Daily expiring-stock alert	Opted-in staff	Viber ("Blu Intrnal") with public list link	Daily schedule; summarizes lots expiring ≤9 months and already-expired, grouped and urgency-sorted	Proactive waste prevention — staff act on near-expiry stock before it is written off
Purchase Requisition approval request	PR Approvers group	Viber ("Blu Intrnal") with no-login tokenized approve/reject link	PR submitted; single-use token	Approvers act from their phone in minutes — no Odoo login — accelerating the PR→PO cycle
Auto-RFQ ready for review	Purchasing Team (or per-orderpoint recipient)	SMS + Viber ("Blu Intrnal")	RFQ created from an approved PR, or orderpoint-driven RFQ flagged to notify	Buyers are pushed straight to new RFQs with vendor and item lines, speeding sourcing
Goods-receipt confirmation	Inventory Team group	SMS (one-liner) + Viber (full per-line breakdown) — full receipts only	PO receipt picking validated to "Fully Received"	Inventory and purchasing are notified the moment stock lands, with received-vs-ordered detail for reconciliation

Overall value: This suite turns routine operational events into instant, mobile-first push notifications — customers stay informed (fewer no-shows, measurable satisfaction), approvals and sourcing happen in minutes via tokenized links instead of logins, and near-expiry stock is acted on before it is lost. Every send is mirrored to the record's chatter or logs, giving management a complete audit trail of who was notified, when, and through which channel — while SMS is reserved for compact alerts and the richer detail rides cost-free on Viber.

Controls, safety & governance

• Test-redirect safety net. A single system setting reroutes every outgoing Viber message to one controlled phone number in non-production environments, so real customers are never accidentally messaged during testing or staging. The reroute is logged with a clear "TEST REDIRECT active" warning, and is disabled simply by clearing the setting — no code change or redeploy required to flip between safe-test and live behaviour.
• Best-effort, non-blocking sends. Messaging is decoupled from core business operations: a failed or errored send is caught, recorded against the message, and never interrupts a repair order, purchase approval, or goods-receipt workflow. Sends are batched and committed incrementally, so a single bad recipient or network timeout cannot stall the rest of the queue.
• Recipient de-duplication & data hygiene. The campaign composer prevents duplicate phone numbers within a campaign and auto-normalises numbers to standard format, reducing the risk of double-messaging a customer or wasting send credits. Each message carries a unique identifier for end-to-end traceability and reliable correlation with delivery events.
• Role- and group-based targeting. Reach is gated by security groups — only Viber Messaging Administrators can run campaigns, and CRM lead alerts fan out only to the curated "Can handle CRM from Viber and AI?" group, which doubles as an editable sales rotation roster. Marketing audiences are built deliberately via partner tags, manual selection, or vetted CSV upload — not broad, uncontrolled blasts.
• Opt-in master switches for AI. Both AI capabilities are off by default and independently controlled in Settings — one switch for AI auto-reply, a separate switch for intent-driven lead creation. Replies can run without lead capture, giving staged, low-risk rollout and the ability to pause either function instantly.
• Branded sub-accounts & sender attribution. Messages route through dedicated, named 8x8 sub-accounts, and each SMS carries a Sender Account tag. Every send is logged with its sending team — a clear audit trail of which department messaged which customer and why.
• Configurable send rate-limit. A central Message Sending Interval (default 5 seconds between messages) acts as a soft throttle to avoid flooding recipients, protect sender reputation with the carrier, and keep send volumes predictable. Batch size is likewise configurable, so send pace can be tuned centrally without code changes.
• Delivery webhooks & closed-loop tracking. 8x8 calls back per message (queued → sent → delivered → read, or failed), updating each recipient's status and auto-closing a campaign once all messages are confirmed delivered. This gives management verifiable proof of delivery, surfaces failures for follow-up, and underpins the AI-chatbot and intent-to-lead audit records.

Adoption & usage (live production data)

Figures pulled directly from the production system on 3 June 2026.

Marketing & AI on Viber

Metric	Value
Viber marketing campaigns run	3 (2 completed) — Dec 2024 to Mar 2025
Campaign recipient messages sent	1,186
Confirmed delivered or read	495 (42%) — of which 390 read (a ~33% read rate)
Failed sends	0
AI concierge status	Live today — ChatGPT gpt-4o; both auto-reply and buying-intent capture currently switched on
Public inquiries auto-answered by the AI	114 (Dec 2024 to Dec 2025)
Curated FAQ entries behind the bot	8
Delivery-tracking events processed	4,520 per-message status callbacks from 8x8

A ~33% read rate on promotional Viber is strong (it compares favourably with typical 20–35% email open rates), and zero recorded failures across 1,186 sends points to clean recipient data on the campaigns run so far. The AI concierge has fielded 114 public inquiries hands-free and remains live with both reply and intent-capture enabled.

Branded SMS

Metric	Value
Total SMS sent to date	16
Sender identities exercised	All five (Coffee, Orders, Service, Acctg, Intrnal)

SMS has been used sparingly — Viber is the primary, richer, lower-cost channel — but every branded sender identity has been exercised and logged.

Automated operational notifications

Metric	Value
Operational Viber notifications issued	308 in the first ~2 weeks (since 21 May 2026)
By sender	Blu Intrnal 288, Blu Service 20

In barely two weeks since launch, the operational layer has already generated 308 staff- and customer-facing notifications — evidence of immediate, heavy uptake across the purchasing, inventory, and service workflows. (This path records dispatch to 8x8; final per-message delivery receipts are captured on the marketing path, not here.)

Current limitations & opportunities

• Inbound two-way handling is limited beyond the AI chatbot. The platform excels at outbound messaging and AI-handled inquiries, but free-form, multi-turn conversations that fall outside the AI concierge's scope are not yet routed to a live-agent inbox. Customers needing complex back-and-forth still rely on existing channels.
• AI reply quality depends on the FAQ training data. The chatbot is only as accurate as the curated FAQ table behind it. Keeping answers current and comprehensive is an ongoing marketing responsibility — gaps or stale entries directly affect the customer experience.
• Reach depends on accurate customer contact data. Both Viber and SMS delivery hinge on having correct, up-to-date mobile/Viber numbers on customer records. Investing in contact-data hygiene directly improves campaign reach and notification reliability.
• Opportunity: turn raw counts into management dashboards. The platform already captures the underlying data — per-message delivery status, AI inquiry volume, campaign reach (see the figures above) — but it is not yet surfaced as live dashboards. Building campaign read-rate, AI-deflection, lead-conversion, and notification-effectiveness views would quantify ROI and guide where to invest next.