Why WhatsApp rewards a tight architecture
WhatsApp is intimate and rule-bound at once: customers expect a personal, near-instant reply, and the platform enforces a 24-hour window and approved templates. A flat prompt ignores both, it sounds generic and it fumbles the window. Architecture is what makes one agent feel personal and stay compliant at the same time.
The four layers
MyChatBot agents stack in layers. System sets a warm, personal identity and guardrails. Role narrows the job, sales, support, order updates, and several can run per number. Memory is the Knowledge Base: Business Domain docs and a Product Feed, so answers are factual. Tools is where it acts, CRM writes, calendar, order creation, template sends. Edit the Feed and answers update; the prompt stays put.
Reading intent and the window
The agent reads intent and the session state together: is this a fresh question inside the window, or a re-open that needs a template? It answers a price question with Agentic Search, confirms an order, or fires the right template to re-engage, always in a personal tone. Same brain, window-aware behavior.
What silently burns credits
The quiet budget killers: re-loading the catalogue per message, re-asking for info already given, and verbose system text on every turn. Fix it with narrow Agentic Search retrieval, CRM-backed memory so context persists across the window, and lean system layers. Architecture is cost control as much as quality.
Designing the handoff and shipping it
Hand-off Control is part of the design: define the stop-word, value threshold and emotional cues that route a chat to a human, with silent hand-off and Flight Control in Telegram. Then ship via the Configuration Wizard, which battle-tests and versions every change, so you tune tone, templates and triggers without ever starting over.