Solution · Business outcomes
LLM chatbots for WhatsApp: answer every customer, day and night
Your customers are already on WhatsApp. An AI agent on the WhatsApp Business API handles the routine — queries, appointments, payments — around the clock, and hands the edge cases to your team with full conversation context.
What this gets you
Staff stop answering the same ten questions
The routine flows — hours, pricing, booking, order status — get handled automatically. Your team handles only what genuinely needs a human, with the conversation history in front of them.
Replies in seconds, at 2 a.m. too
Webhook ingestion and async processing (Redis + Celery) mean a customer gets a reply in seconds, whether it's noon or the middle of the night — without the reply blocking on the model.
Context that survives the conversation
Deliberate context management across turns — what to carry forward, summarise, or drop — keeps long conversations useful without blowing the token budget.
Documents stop getting lost in group chats
Ingestion pulls shared media into Google Drive, with a recovery job that backfills anything missed during downtime — built and deployed at Zuneko Labs.
How it works
1. WhatsApp Business API webhook
Meta's platform sends every inbound message to your webhook. Provider event IDs are stored with unique constraints so retries never double-process a message.
2. The agent replies asynchronously
Message processing, AI inference and database writes run on Celery workers, so the webhook returns instantly and the customer never sees a timeout.
3. Structured flows where it matters
For anything with money or commitments, the critical steps are structured flows, not free-form generation — with a human hand-off path and logging of what the agent did.
4. Business-initiated messages follow Meta's rules
Approved templates, and free-form replies within the window after the customer's last message. The design works inside those rules, not around them.
Frequently asked questions
Can the agent take payments?
Appointment and payment flows are built as structured steps, not free-form chat — the model handles the open-ended parts and hands off to deterministic flows where money is involved.
What happens when the agent doesn't know the answer?
It says so and hands off to a human with the full conversation context. Agents are also configured to improvise only within defined limits, with logging of everything they do.
Do I need Meta's business verification?
Yes — WhatsApp Business API access and pricing are controlled by Meta, and business-account approvals are outside any developer's control. I'll guide the setup, but the approval is yours.
Hosted or self-hosted LLM?
Hosted APIs (OpenAI, etc.) are fastest to ship. Self-hosting LLaMA gives more control over data and cost at volume, but you run the GPUs. The choice depends on your volume and data sensitivity.
Your question not here? Ask directly
Want this built for your business?
Book a free 15-minute call below — I'll tell you honestly whether this fits your situation, and what it would take.
How I work with clients