A WhatsApp AI agent for sales, support and your CRM.
Not another canned-response bot. Two smart paths in one channel: a discovery agent for new leads, and a support agent for active clients that can open tickets, read screenshots and voice notes, send updates, and keep the full context inside your system.
Alcyone14 Agent
Live routing
One WhatsApp. Two different jobs.
The important part is not that the bot answers. The important part is that it knows who it is answering and what it is allowed to do.
Sales agent for leads
Collects pain, understands the current stack, explains the Skin in the Game model, and books a call only when the conversation fits.
Support agent for clients
Opens tickets, answers how-to questions inside the client system, stores media, and updates the client when work is resolved.
What the agent actually does.
This is not AI for the sake of AI. These are real business actions handled from WhatsApp.
Lead discovery
Asks which tools are driving the business crazy, what breaks today, and what needs to become one system before the engineer call.
Booking and reminders
Books the discovery call, sends confirmation, reminds the lead, and supports direct WhatsApp cancellation or rescheduling.
Tickets from WhatsApp
Bug, feature, request or question becomes a ticket with project, priority, category and context.
Screenshots and voice notes
The client sends an image or audio. Media is stored, shown in chat, and attached to the ticket.
Invoices and status updates
Clients receive approved template messages for invoices, resolved tickets and new releases.
CRM-aware answers
The agent uses screen maps and system knowledge to explain where a button lives, how to reach a page, and what changed.
Redaction before AI
Names, phones and sensitive details become placeholders before model calls and are restored only for the final response.
Human escalation
If there is a complaint, missing knowledge or risk, the agent does not invent. It routes to the team with context.
What happens behind one message.
For this to be a real product, every message follows a clear path: identify, redact, decide, act, document.
Identify
The phone number is checked against active contacts. New leads go to sales; active clients go to support.
Context
The system loads project, CRM, open tickets, recent releases and escalation rules.
Protect
Sensitive details are replaced with placeholders before AI so unnecessary personal data does not leave the system.
Act
The agent answers, books, opens a ticket, stores media, or sends a template update.
Document
Everything is saved in the CRM and ops panel so the engineer sees what happened without asking again.
The part clients feel fastest.
A client does not need to know where to open a ticket. They send a message like they normally do, and the system organizes the work behind it.
Screenshot
The client captures the bug. The image is stored and shown inside the ticket.
Voice note
The client explains instead of typing. Transcription enters the ticket with the original media.
Linear + CRM
The ticket connects to the right project, with a Linear link when engineering work is needed.
Update back
When a ticket is resolved or a release is published, the client gets a clean update.
Why this is not another chatbot.
Regular chatbot
Answers prepared questions, breaks when context matters, and is usually disconnected from the real work.
Alcyone14 agent
Knows if this is a lead or client, connects to CRM, opens tickets, stores media, and escalates when needed.
Questions buyers ask.
If WhatsApp already runs the business, turn it into the system.
Short call. We will check whether you need the sales agent, the customer agent, or both.
Send a message