Autonomous Sales Agent
An AI sales agent that moves leads from first message to qualified opportunity — reading intent, pricing dynamically, and alerting a human the moment a lead turns hot.
At a glance: Ethereal built an AI sales assistant for a travel brand that chats with customers on WhatsApp and Telegram day and night, answers their questions, works out the right price, and pings the owner the moment a buyer is ready to purchase. It passed all 62 of its tests and never lets an overnight lead go cold.
The Problem
A travel-commerce brand was losing warm leads to manual WhatsApp/Telegram handling: no after-hours response, inconsistent pricing calculations, and no systematic follow-up.
Leads arriving overnight on WhatsApp and Telegram went unanswered until someone was back at a desk.
Every inquiry was triaged, priced, and followed up by hand across two messaging channels.
Ad-hoc quoting and follow-up produced uneven pricing and leads that quietly went cold.
The Solution
We implemented an autonomous agent that extracts intent and handles multi-turn customer conversations. A dynamic pricing engine resolves seasonal multipliers and discount conflicts; RAG over the client's knowledge base answers product questions; and the full lead lifecycle is automated — new → warm → hot → contacted — with hot-lead alerts to the owner and automatic re-engagement of stale leads. An admin portal adds Excel upload, audit logs, and analytics.
Tenant Config
Each client's packages, pricing rules, and knowledge base are declared, not coded.
RAG Grounding
FAISS vector search grounds every answer in the client's own knowledge base.
Session Memory
Redis-backed multi-turn sessions merge intent across a conversation.
Agent Engine
One runtime handles intent extraction, dynamic pricing, tool use, and reply generation.
logic tests passing
Seasonal multipliers and discount-conflict resolution resolve to one defensible quote.
Stale leads are automatically re-approached before they lapse.
Prompt-injection attempts are caught before they reach the model or tools.
Cross-tenant access is blocked so no client can see another's data.
Per-contact limits absorb floods and abusive traffic at the edge.
Inbound WhatsApp and Telegram payloads are signature-verified before processing.
Every pricing change and lead action is written to an immutable audit trail.
The agent handles, prices, and routes leads around the clock without a human in the loop.
The Outcome
Fully automated 24/7 lead qualification. Verified through 62/62 passing logic tests, with rate limiting, prompt-injection prevention, and cross-tenant isolation.
Built with
06 technologiesHave an agentic ai problem like this?
Tell us what you're building — we'll tell you exactly how we'd ship it.
From the archive
An AI home-advisor for real estate that qualifies buyers, scores its own confidence, and escalates hot leads over WhatsApp — tuned by a three-layer evaluation suite.
Three parallel AI agents that audit websites, product flows, and user experiences from multiple perspectives at once — with video evidence.
Describe a multi-leg options strategy in plain English — the copilot constructs it, validates it against risk rules, and executes across accounts.
An agent that sanitizes enterprise change requests, scores compliance, and plugs straight into the ITSM workflow — with a full audit trail and zero PII leakage.