DEV Community: Indra Gunanda

Why We Built a CRM Around WhatsApp Instead of Email

Indra Gunanda — Mon, 27 Jul 2026 08:55:37 +0000

Why We Built a CRM Around WhatsApp Instead of Email

In Southeast Asia, email open rates hover around 15%. WhatsApp message read rates? North of 90%.

That single stat explains why we built Zetta CRM and Hallo Zetta as WhatsApp-native tools instead of bolting messaging onto yet another email-first CRM.

This article walks through the architectural and product decisions behind building AI-powered customer communication on WhatsApp — and what we learned shipping it to real teams.

The Problem With "WhatsApp Integration"

Every CRM claims WhatsApp support. Here's what that usually means:

A third-party plugin that syncs messages with a 5-minute delay
A sidebar panel showing chat history you can't reply from
Template-only outbound that feels robotic to customers

None of these solve the actual workflow: a team of 3-10 people sharing one WhatsApp number, handling hundreds of conversations daily, across DMs and groups, with photos, voice notes, and documents flying around.

That's what Zetta CRM was designed to handle from day one.

Architecture: WhatsApp as Primary, Not Peripheral

Our stack treats WhatsApp as the primary communication channel:

WhatsApp (multi-device) → Zetta Gateway → Message Router
                                              ↓
                              ┌────────────────┼────────────────┐
                              ↓                ↓                ↓
                        AI Agent         Team Inbox        Integrations
                     (Hallo Zetta)     (shared view)    (Calendar, Sheets,
                                                         Notion, Webhooks)

Key design decisions:

Native multi-device support: Connect via QR scan, no WhatsApp Business API dependency
Full message types: Text, images, documents, voice notes, contacts, locations
Group awareness: The system understands group dynamics — who said what, reply threading, @mentions
Real-time sync: Sub-second message delivery to the shared inbox

AI Layer: Hallo Zetta

Hallo Zetta is the AI agent layer on top of the CRM. It's not a chatbot — it's an AI teammate:

What makes it different from chatbots:

Chatbot	Hallo Zetta
Script-based, breaks on unexpected input	LLM-powered, handles natural conversation
Replies to everything in groups	Silent until @mentioned or quoted
Generic answers	Grounded in YOUR knowledge base
No handoff	Seamless human takeover per-conversation
Text only	Photos, files, voice notes

The knowledge base workflow:

Upload your docs (product info, pricing, FAQ, policies)
Test in the playground — see how AI answers before going live
Publish to the agent
Monitor and improve based on real conversations

Every answer traces back to source material. No hallucination. No made-up pricing. No "I think the delivery time is..." guesses.

Group Intelligence

This is where most WhatsApp automation tools fail catastrophically.

Groups are noisy. A reseller group might have 200 messages/day, of which maybe 5 are actual questions for the brand. A naive bot replying to everything gets muted or kicked within hours.

Hallo Zetta's group behavior:

Default: silent — never interrupts human conversation
Activates on: @mention or direct quote of the bot's message
Replies in-thread: Responds to the specific message, not the whole group
Context-aware: Reads the conversation thread before answering

This is a product decision, not a technical limitation. Trust in groups is earned by knowing when NOT to speak.

Integrations That Move Work Forward

A conversation often triggers work elsewhere. Instead of copy-pasting between apps:

Google Calendar: Schedule meetings directly from chat
Google Sheets: Push qualified leads to a sheet the sales team already watches
Notion: Sync knowledge base from Notion pages
Zapier & Webhooks: Connect to 6,000+ apps or your own backend

The webhook layer means any system with an API can react to conversation events: new lead captured, label applied, handoff triggered.

Results in Production

Teams using Hallo Zetta report:

85% of routine questions handled by AI without human intervention
Response time dropped from 4+ hours to under 30 seconds
Zero noise complaints in group deployments (the silent-by-default pattern works)
Team efficiency up 3x — same team handles 3x the conversation volume

The Zetta CRM Ecosystem

Hallo Zetta is part of the broader Zetta CRM platform:

Contacts & Labels: Unified contact database with custom labels and segments
Team Inbox: Shared workspace with role-based access
Analytics: Conversation metrics, response times, resolution rates
Multi-number: Connect multiple WhatsApp numbers to one workspace
API & MCP: Programmatic access for developers building on top

Built by Cipta Dusa in partnership with Incredible Zetta.

Try It

If your team handles customer communication on WhatsApp and you're tired of:

Personal phones as the primary tool
Same questions answered 50 times a day
No visibility into what your team is saying
Conversations lost when someone leaves

Check out Hallo Zetta. Connect a number in 2 minutes, upload your docs, and see the AI handle the first conversation.

Part of the Zetta CRM family. Built by Cipta Dusa — software development for teams that move fast.

How We Ship Websites in 48 Hours Without Cutting Corners

Indra Gunanda — Mon, 27 Jul 2026 08:54:52 +0000

How We Ship Websites in 48 Hours Without Cutting Corners

Everyone says "fast delivery" until you ask them what fast actually means. For us at Cipta Dusa, it means a company profile website — designed, built, and live — in two working days.

Not a template slap. Not a drag-and-drop page builder with your logo swapped in. A real, hand-coded site with proper SEO structure, responsive design, and performance that scores 90+ on Lighthouse.

Here's how we do it without burning out or shipping garbage.

The System Behind the Speed

Speed without a system is just chaos. We built ours over dozens of projects:

Day 1: Discovery + Design

30-minute intake call: business type, target audience, must-have pages, brand assets
Component selection from our battle-tested design system (not templates — reusable, customizable blocks)
Client gets a preview link by end of day

Day 2: Build + Deploy

Full responsive implementation
SEO meta tags, Open Graph, structured data (Organization schema, breadcrumbs)
Performance optimization: lazy loading, WebP images, minimal JS
Deploy to edge CDN, SSL, custom domain
Client walkthrough + handoff

Why This Works for Indonesian SMEs

Most small businesses in Indonesia don't need a 3-month web project. They need:

A professional online presence that loads fast on mobile
Something they can show investors, partners, or customers
Google-indexable pages with proper local SEO

That's exactly what we deliver. No scope creep, no endless revision loops, no "we'll get back to you next week."

Beyond Websites: Full-Stack Development

Websites are our entry point, but Cipta Dusa builds the full stack:

Custom web applications — dashboards, internal tools, SaaS products
Android apps — native and cross-platform
AI chatbots — WhatsApp, Telegram, and website integrations powered by LLMs
DevOps setup — CI/CD pipelines, cloud deployment, monitoring

Every project follows the same principle: ship fast, ship quality, keep the client in control.

Our Tech Stack

We're opinionated about tools because opinions save time:

Frontend: Next.js, Astro, Tailwind CSS
Backend: Go, Node.js, Python
Infrastructure: Cloudflare Workers, Vercel, Docker, GitHub Actions
AI: OpenAI, Anthropic, local models via llama.cpp
CRM & Automation: Zetta CRM (our own)

Results

In the last 6 months:

40+ websites delivered
Average delivery time: 2.1 days
Client satisfaction: 4.8/5
Zero missed deadlines

If you need a website that's live before the weekend, or a custom app that doesn't take 6 months, check out our portfolio at ciptadusa.com.

Built by the team at PT Cipta Dua Saudara — software development partner for startups and SMEs in Indonesia.

Building an AI Support Inbox for How WhatsApp Teams Actually Work

Indra Gunanda — Mon, 27 Jul 2026 07:23:34 +0000

Building an AI Support Inbox for How WhatsApp Teams Actually Work

Most customer-support software starts from an email-shaped assumption: messages arrive one at a time, belong to one person, and wait patiently in a queue.

WhatsApp does not work like that.

A customer may send a photo, then a voice note, then three short follow-ups. A team may need to jump between direct messages and busy groups. Someone has to take over a conversation without asking the customer to repeat everything. And when automation answers at the wrong moment, it does not feel helpful — it feels intrusive.

That was the problem behind Hallo Zetta: build a shared WhatsApp customer-care inbox where AI can handle routine work without pretending humans are no longer needed.

Start with WhatsApp behavior, not chatbot behavior

It is easy to describe the first version of an AI support product:

receive a message,
send it to a model,
return an answer.

That loop is not enough for a real support team. It ignores context, ownership, files, group dynamics, and the moment a conversation should move to a human.

So we treated WhatsApp as the primary working environment, not as one more integration tab inside a generic CRM. The product needs one shared inbox for the team, persistent contact context, and a conversation model that can carry AI and human work in the same thread.

The design target was not “make the bot reply more.” It was “help the team move the conversation forward.”

Groups change everything

Groups are where a naïve support agent becomes noise.

A useful AI agent cannot reply to every message in a reseller group, customer group, or internal coordination thread. It needs to recognize that a group has its own social rules. Most messages are not requests for the agent. Many are conversations between people who do not need an interruption.

Hallo Zetta is built to support group-aware workflows. The agent should participate only when invited into the interaction, while the rest of the group stays human. That decision is product behavior, not a cosmetic setting: it protects trust in a channel where one bad automation pattern can make a team stop using the tool.

This also changes how we think about context. A reply is not only an answer. It needs to be tied to the right message, the right group, and the right moment.

Ground AI answers in team knowledge

Another failure mode is confident automation with no source of truth.

A support agent should not invent pricing, policy, delivery, or product answers. The useful version of AI is constrained by material the team has reviewed.

Hallo Zetta uses a knowledge-base workflow and a playground so teams can test answers before the agent is live. That creates a practical loop:

publish team knowledge,
test likely customer questions,
inspect weak or incomplete answers,
improve source material,
deploy with more confidence.

The goal is not to remove judgment from support. The goal is to make good judgment reusable.

Handoff is a core feature

“Human handoff” often sounds like an edge case. In a real inbox, it is central.

Some questions need empathy. Some need approval. Some need access to systems or decisions that should never be automated. A good system makes those transitions boring: a teammate takes over in the same conversation, sees the context, and continues without forcing the customer through a second introduction.

That is why the AI-first model matters. AI handles routine first replies and repetitive questions. Humans retain control of exceptions, escalation, and relationship-building.

AI-first is not AI-only.

An inbox must handle more than text

Support work includes photos, documents, voice notes, and links. A shared inbox that only handles clean text messages forces teams back to personal phones for everything else.

Hallo Zetta is designed around the full shape of WhatsApp conversations: media, files, voice, contact context, and team ownership. Each capability seems small in isolation. Together, they decide whether a tool can replace a patchwork of personal devices and workarounds.

Integrations should move work, not create more tabs

Support conversations often trigger work outside the inbox: schedule a follow-up, record a lead, update a workspace, or send an event to another system.

That is why Hallo Zetta connects with tools teams already use, including Google Calendar, Google Sheets, Notion, Zapier, and webhooks. The point is not a long integration checklist. The point is to keep a customer conversation from becoming copy-paste work across five products.

What we are building toward

The hardest part of customer support is not generating a sentence. It is knowing what to say, when to say it, when not to say anything, and when to involve a person.

That is the product direction behind Hallo Zetta: one WhatsApp-native workspace where a team and an AI agent can work together without losing context or control.

If your support team lives in WhatsApp, try it at hallo.zettacrm.com.

We Built Hallo Zetta Because We Were Tired of Watching Teams Answer WhatsApp on Personal Phones at Midnight

Indra Gunanda — Mon, 06 Jul 2026 12:45:34 +0000

The story behind why we built a WhatsApp CRM that actually understands how WhatsApp works.

There's one scene I can't get out of my head.

A friend's desk. She runs an online store. On it sat three phones. Not for show. One for customer service, one for the admin, one for the number that was "just for resellers." All three buzzing, nonstop. And there she was, eleven at night, still replying to messages one by one, sighing: "It's the same questions over and over. But if I don't reply, they'll go to the competitor."

That's not a rare case. That's the normal state of things for thousands of businesses.

We all know one thing CRM software rarely admits: customers here don't live in email. They live on WhatsApp. They ask about prices on WhatsApp, complain on WhatsApp, close deals on WhatsApp, even ask for warranty support on WhatsApp. But the teams handling all of it? They use personal phones. No records, no context, no way to help each other when one person is drowning.

Hallo Zetta was born out of that.

What Frustrated Us About the Existing Tools

Before building our own, of course we looked. Surely someone had solved a problem this simple?

Turns out what existed fell into two camps, and both were maddening.

Camp one: dumb auto-reply bots. Type "hi," get a template. But the moment a customer asks something slightly off-script, the bot freezes. It actually makes customers angrier, because it feels like talking to a wall.

Camp two: bloated CRMs. Loaded with features, dashboards full of charts, but WhatsApp is bolted on as one small tab. As if WhatsApp were an afterthought, not the main battlefield. For most of our customers, WhatsApp is the battlefield.

Nothing fit. So we decided to build it ourselves.

The Hard Part Isn't "AI Can Reply to Messages"

Let me be honest about this. Bolting AI onto WhatsApp is easy. Anyone can wire GPT to a webhook and ship it overnight. If that were the whole goal, this article wouldn't need to exist.

The hard part, the thing that made us rethink everything again and again, is groups.

Think about it. How much real business actually happens in WhatsApp groups? Reseller groups, VIP customer groups, team coordination groups. A huge amount. But an AI that chimes in on every group message isn't helpful, it's spam. Picture a group in the middle of a lively conversation, and suddenly there's a bot butting in over and over. You'd get kicked out in seconds.

So we made a rule the bot genuinely respects:

Inside a group, the bot stays silent. Completely silent. Until it's mentioned (@) or someone quotes its message. Only then does it answer, right in the correct thread, quoting the message it's responding to.

No unnecessary noise. It behaves like a teammate who knows their place, not a chatbot that can't stop talking. Sounds trivial, but this is what makes the bot genuinely usable in groups, not just in private chats.

AI First, But Not AI Only

The second principle we refused to compromise on: control stays with the team.

The AI in Hallo Zetta does answer first. Routine questions, opening hours, shipping rates, payment methods, answered instantly, 24/7. But it only answers from the knowledge base you publish yourself. It doesn't make things up. If you didn't write about it, it won't invent an answer.

And your team is always one tap away. A conversation that needs a human touch? Just switch the AI off on that chat, take over in the same thread, context and media all intact. No app-switching, no asking "what did we talk about again."

That's our philosophy: AI-first, not AI-only. Let AI handle the routine. Give the human-needed conversations a human. You decide where the line is.

The Boring Work We Did Behind the Scenes

To make the inbox "just work," there was a lot of tedious work we had to sort out first.

Real WhatsApp is messy. There are DMs, groups, mentions, quoted replies, photos, files, even voice notes. There's the headache of phone-number vs LID identity. There's media that has to be stored securely. We handled them one by one, because these are exactly the details that usually make other tools "look like they work" until you actually use them for real.

There's also a playground. Before your agent talks to real customers, you can test its answers there first. Match them against the knowledge base, fix the odd ones, then publish. No embarrassing surprises in front of a customer.

It's Live Today

Hallo Zetta is live today. It's part of the Zetta CRM family, built together by Incredible Zetta and Ciptadusa.

Getting started is simple:

Connect your WhatsApp number (scan a QR or use a code).
Teach the AI by publishing your docs to the knowledge base.
Run the inbox and let AI handle the routine while your team steps in when needed.

You can try it right now, connect a number, publish your docs, and see for yourself how it handles a real conversation.

👉 hallo.zettacrm.com

Closing

We didn't build this to jump on the AI trend. We built it because we were tired of watching people we know scramble through WhatsApp with manual workarounds that wear them out.

If your team is still juggling three phones on a desk, or still answering the same questions until midnight, this might be for you.

Try it, and tell us what's missing. We're listening.

Written by the Hallo Zetta team. If this story landed, give it a clap 👏 and share it with a friend still fighting their WhatsApp inbox