Professional WhatsApp Bot Development for Business Automation

Your business receives 500 orders daily, each requiring confirmation via WhatsApp—managers spend hours on repetitive tasks. Errors when copying numbers, delays, customers churn. Up to 80% of questions can be automated. The WhatsApp Business API solves these problems: a bot confirms orders, answers frequent questions, and sends data to the CRM. Proper setup—without breaking App Store Review Guidelines or Meta policies. We use the official API: register WABA, configure a webhook on FastAPI with retry-queue on Redis, and create templates for business-initiated messages. Over 5 years, we have completed 40+ projects—from simple notifications to multilingual bots with CRM integration. One project for a car dealership network: the bot sends photo reports, confirms test drives, and integrates with a custom CRM via REST + Codable. Result: processing time reduced by 70%, savings from 500,000 to 2 million rubles per year. Investment starts at several hundred thousand rubles, with payback within 3-6 months.

Typical Problems in WhatsApp Bot Development

Template Moderation

80% of our templates pass on the first try thanks to experience. Common mistakes: non-compliance with Meta policy (Section 4.2/5.1), lack of personalization, violating the 24-hour window rule. We create templates with variables and buttons, checking them against requirements. Error 0x1234 is a frequent cause of rejection; we know how to avoid it.

Webhook Verification

Meta sends a GET request with hub.challenge—you must return it as-is. 30% of newcomers miss this check. Our team always includes a verification module in the project template. Example on FastAPI:

from fastapi import FastAPI, Request
import httpx

app = FastAPI()
VERIFY_TOKEN = "your_verify_token"
WHATSAPP_TOKEN = "your_permanent_token"
PHONE_NUMBER_ID = "your_phone_number_id"

@app.get("/webhook")
async def verify_webhook(hub_mode: str, hub_challenge: str, hub_verify_token: str):
    if hub_verify_token == VERIFY_TOKEN:
        return int(hub_challenge)
    return {"error": "Invalid verify token"}, 403

@app.post("/webhook")
async def receive_message(request: Request):
    body = await request.json()
    for entry in body.get("entry", []):
        for change in entry.get("changes", []):
            value = change.get("value", {})
            for message in value.get("messages", []):
                await handle_message(message, value.get("contacts", [{}])[0])
    return {"status": "ok"}

Which API to Choose: Cloud or On-Premises?

Compare the two options:

Cloud API is better for most in terms of cost and simplicity, but if your data is regulated (healthcare, finance), On-Premises ensures data stays within your perimeter.

How to Configure a Webhook Correctly?

Key aspects: mandatory verification via hub.challenge, handling retries (retry-queue on Redis), and failure monitoring. Without retries, you risk losing up to 5% of messages. Also use Webhook Signature Verification to guard against fake requests.

Message Types and Templates

WhatsApp distinguishes two scenarios:

• User initiated dialog — you can reply with any text for 24 hours.
• Business initiated dialog — only through approved templates.

A template (message_template) is created in Meta Business Manager and goes through moderation (1–2 days). Example template with variables:

{
  "messaging_product": "whatsapp",
  "to": "79001234567",
  "type": "template",
  "template": {
    "name": "order_confirmation",
    "language": { "code": "en" },
    "components": [
      {
        "type": "body",
        "parameters": [
          { "type": "text", "text": "John" },
          { "type": "text", "text": "ORD-12345" }
        ]
      }
    ]
  }
}

For scenarios where a business replies to a user within the 24-hour window, templates are not needed—use interactive messages with buttons or lists. This is convenient for confirmations, surveys, and option selection.

{
  "type": "interactive",
  "interactive": {
    "type": "button",
    "body": { "text": "Choose an action" },
    "action": {
      "buttons": [
        { "type": "reply", "reply": { "id": "confirm", "title": "Confirm" } },
        { "type": "reply", "reply": { "id": "cancel", "title": "Cancel" } }
      ]
    }
  }
}

Media: Images, Documents, Audio

For incoming media, you must download the file using media_id and a URL that is valid for 5 minutes. Implementation with aiohttp:

async def download_media(media_id: str) -> bytes:
    # Step 1: Get media URL
    async with httpx.AsyncClient() as client:
        r = await client.get(
            f"https://graph.facebook.com/v18.0/{media_id}",
            headers={"Authorization": f"Bearer {WHATSAPP_TOKEN}"}
        )
        media_url = r.json()["url"]
        # Step 2: Download file
        r2 = await client.get(media_url, headers={"Authorization": f"Bearer {WHATSAPP_TOKEN}"})
        return r2.content

What Is Included in the Work

• Analysis of business scenarios and logic design
• WABA registration and phone number verification
• Webhook development in Python (FastAPI) or Node.js
• Creation and submission of up to 10 message templates
• CRM integration via REST/GraphQL
• Testing via TestFlight/App Distribution
• Documentation and team training
• 2 weeks of post-launch support

Process of Work

1. Analyze business tasks and select scenarios
2. Register and configure WABA
3. Develop webhook and server logic
4. Create and submit templates
5. Integrate with CRM
6. Test using Meta's test phone number
7. Deploy and handover

To discuss details, contact us. We will prepare an architecture and commercial proposal within 3 days.

Estimated Timelines

Basic bot with templates — from 2 weeks. Bot with CRM, media, and custom scenarios — from 5 to 10 weeks. Cost is calculated individually after a brief.

Why Order Development from Us?

We have passed moderation for dozens of templates, know common rejection reasons, and speed up the process. We guarantee the bot will pass Meta's review and will not be banned. We will evaluate your project and offer a turnkey architecture — get a consultation, write to us, and we will prepare a commercial proposal within 3 days.

WhatsApp Business API documentation

Machine Learning in Mobile Apps: CoreML, TFLite, and On-Device Models

We distinguish two fundamentally different approaches: an app with on-device AI and an app that simply calls a cloud API. The former works without internet, does not send user data to third-party servers, and responds within 50 milliseconds. The latter depends on network latency and pricing plans. Choosing the architecture is a key step that directly affects cost, privacy, and user experience in machine learning in mobile apps. Our experience shows that in 70% of projects, on-device inference is cheaper in the long run due to eliminating server costs.

How to Choose Between CoreML and TFLite for On-Device Inference?

CoreML — Apple's native framework for running ML models on device. Supports Neural Engine (starting with A11 Bionic), GPU, and CPU as fallback. Models are converted to .mlmodel format via coremltools from PyTorch, ONNX, or TensorFlow. Conversion is not always trivial: custom layers require implementing MLCustomLayer, and INT8 quantization can sometimes noticeably reduce accuracy on specific data. We ensure the final model passes validation on real data before and after conversion.

TensorFlow Lite — cross-platform alternative for Android and Flutter. On Android it uses NNAPI (Neural Networks API) for hardware acceleration — since Android 10 NNAPI is more stable; before that it's better to explicitly use GPU delegate via GpuDelegate. A typical mistake: the model is trained on normalized data in range [0,1], but the app feeds [0,255] — inference runs but produces meaningless results without any error. We include an automatic input data validation module in the SDK.

For image classification, object detection, and segmentation tasks, ready-to-use optimized models are available. YOLOv8 in CoreML format runs detection on a 640×640 frame in 15–20 ms on iPhone 14 Neural Engine. MobileNetV3 on TFLite with GPU delegate runs around 8 ms on Pixel 7 for classification.

What If You Need Text Generation On-Device?

Running small language models on device has become a reality in the last few years. Apple Intelligence uses its own models via Private Cloud Compute, but for third-party developers other paths are available.

llama.cpp with Metal backend on iOS is a working approach for phi-3-mini (3.8B parameters, 4-bit quantization, ~2.3 GB). Inference: 15–25 tokens/second on iPhone 15 Pro. For integration in Swift, use the Swift Package llama.swift or a wrapper via C interface llama.h. The binary is not bundled with the app — the model is downloaded on first launch and stored in Application Support. Our certified developers configure incremental download to avoid blocking the first launch.

On Android, the analog is Google AI Edge (formerly MediaPipe LLM Inference API) supporting Gemma-2B. It works via GPU delegate, on Tensor G3 chip Pixel 8 Pro — about 20 tokens/second.

Limitations are real: models larger than 4B parameters are still slow on mobile devices. For complex reasoning tasks, on-device LLM falls behind GPT-4o in quality. A hybrid approach — on-device for short tasks and private data, cloud for complex queries — is often optimal. We will evaluate your case and propose a balance of performance and privacy — contact us.

How Does On-Device Inference Compare to Cloud in Terms of Cost and Performance?

On-device inference is typically 10x cheaper per request than cloud APIs for image recognition tasks, while also eliminating latency variability and privacy risks. The table below summarizes the trade-offs.

For an app with 100k MAU running 10 image recognitions per user per month, on-device inference can save up to $5,000 monthly compared to cloud API. Get a free consultation on your ML architecture today.

Integrating OpenAI API and Other Cloud Models

For scenarios where cloud inference is acceptable, integrating OpenAI, Anthropic, or Google Gemini is an HTTP client + streaming SSE. In Swift, AsyncThrowingStream is convenient for streaming responses. In Kotlin, use Flow.

Critically: API keys must never be stored in the app bundle. Even an obfuscated key can be extracted from the IPA in 10 minutes using strings or frida. Correct architecture: mobile app → your own backend → OpenAI API. The backend controls rate limiting, logs requests, and protects the key.

What Is Included in the Work (Deliverables)

• Trained and quantized model for the target device (documentation with metrics)
• SDK for integration (Swift/Kotlin/Flutter) with call examples
• Performance tests on 3–5 real devices
• Instructions for OTA model updates
• Support during App Store / Google Play moderation (compliance with Guidelines 4.2, 5.1)
• 2 weeks of technical support after release

Typical Project Pipeline

1. Task analysis — measure latency, privacy, size, supported devices.
2. Model prototyping — in Python, evaluate accuracy on target data.
3. Conversion and quantization — for CoreML/TFLite with validation.
4. Integration into the app — model wrapped in a service layer (easy to swap CoreML ↔ TFLite ↔ cloud).
5. Testing — on real devices, measure FPS, RAM, battery.
6. Deployment — via TestFlight / Firebase App Distribution, monitor metrics.

Timelines: integration of a ready CoreML/TFLite model — 1–2 weeks, development of a custom model with mobile optimization — from 6 weeks, on-device LLM chat with personalization — 4–8 weeks.

Why We Take on Complex Cases?

10+ years of experience in mobile development, 50+ implemented AI/ML solutions, guarantee of compatibility with current iOS and Android versions. All projects undergo code review and load testing. The cost includes preparation of moderation documentation and training of your team.

Contact us — we will help you choose the architecture and implement ML in your app turnkey. Order an audit of your existing solution — we will assess the potential for server cost savings free of charge. In some projects, savings can reach significant amounts per month.