Skip to main content

Chatarmin AI - WhatsApp AI Chatbot

How to set up and use the Chatarmin WhatsApp AI chatbot for customer support and sales automation

Updated this week

Chatarmin AI

You can integrate your own Chatarmin AI bot into Chatarmin to help solve your incoming customer requests via WhatsApp. And, automate leadgen and lead-enrichment based on logics and smart-splits you train your chatbot to execute. After buying your Smarter, Genius, Master or bigger subscription you can get access to Chatarmin's AI and thus start automating your customer support and WhatsApp Flows.
​Feel free to inquire at our Team here, to get your Chatarmin AI quote.

Train and setup your Chatarmin AI Chatbot

  1. Go to the “AI” tab in your Chatarmin Dashboard.

  2. Define your instructions
    Use the Master Prompt feature to customize your chatbot’s tone, personality, behavior, and logic. This lets you shape how your bot sounds and reacts. We recommend making the instructions highly specific to your business case and datasets.
    Here is an instructions template that you can customize however it fits best

    for your needs: Instructions Template

  3. Add integrations
    Chatarmin AI currently supports integrations with:

  4. Add your data sources
    You can train your bot with a wide range of sources:

    • 🔧 Files ( Documents with instructions or product info)
      ​Example: PDFs, guides, User manuals, product brochures.

    • 🔧 Q&A content (FAQs that answer common customer questions.)
      ​Example: “How long does delivery take?”, “What is the return policy?”

    • 🔧 Websites (URLs or sitemaps to guide the customer)
      ​Example: Blog-page, Home-page, Product-page

    • 🔧 Product catalogs (e.g., WhatsApp product listings to show and suggest products)
      ​Example: Recommendation of best-selling shoes with image & current price

    • 🔧 Salesforce CRM (optional - activate to train your bot on sales, support, and CRM data)

  5. Set the retraining frequency
    For optimal performance, we recommend setting your retraining frequency to every 24 hours, so the chatbot continues learning automatically from your updated data.

  6. Test your bot before go-live

    Manual testing is essential. You can simulate conversations in the dashboard and adjust responses as needed.

  7. Trigger the AI in your chat flow

    Upon chat start, choose “AI” as the action in your flow. This ensures your bot responds based on the instructions, sources, and integrations you configured.


    ​Example of a client using Chatarmin AI Chatbot for his fully automated client support. Integrating the Chatarmin AI chatbot with zendesk, billbee, and Shopify.

Retrain your Chatarmin AI Chatbot

Whenever you add relevant sub-pages to your website or update your website sitemap, it’s important to manually retrain your Chatarmin AI chatbot to ensure it stays accurate and informed.

The same applies if you create new guides, manuals, or prompts - always make sure to feed this content into the chatbot and press the “Retrain” button.

⚠️ Important:
Make sure to retrain the chatbot within the correct sub-source page.
For example, if you're updating website content, first navigate to the “Websites” section, and then press “Retrain Chatbot” there.
This ensures that the AI is only trained on the content relevant to that source.

Repeat this for all other training sources (e.g., documents, support pages, etc.) - each source must be retrained individually.

Smart Split

Smart Split is a feature in our Chatarmin AI chatbot designed to forward the conversation to your team or guide the user further through a flow. For example, if a customer sends images, videos, or any kind of input the bot can't handle, Smart Split can redirect the chat to a human agent.
​

It can also trigger when the customer seems frustrated or when you prefer not to handle certain topics - like pricing or return requests - via the chatbot. This ensures a smooth user experience and lets your support team take over when needed.

Improve Answer Manually

After conversations are completed, you can review and refine the AI’s responses via the “Improve Answer” tab in your AI dashboard.

Source: Chatarmin, Q3 2024

You’ll find your answer improvement history and documentation in the improve answer tab, helping you track all changes over time: AI-dashboard

Source: Chatarmin, Q3 2024

Chatarmin AI Product Catalog Integration

You can integrate your product catalog into Chatarmin AI, allowing your bot to:

  • Search products

  • Recommend items

  • Guide customers through purchases directly via WhatsApp

Source: Chatarmin, Q3 2024

Read more about our WhatsApp catalog feature here.

Integrate Chatarmin AI Chatbot with your Customer Support Tool

You do have the chance to enable our Chatarmin AI to integrate with your customer support tool, such as:

  • Freshdesk

  • Zendesk

  • Gorgias

  • DIXA

  • Intercom

  • HubSpot

  • and several others

Check out this video, to see how it works:

Source: Chatarmin, Q3 2024

Storefinder: Send Location & present nearby stores!

If your client is looking for the nearest store, you can enable this feature in your AI bot to offer a seamless experience.

Here’s how the user journey works:

Source: Chatarmin, Q3 2024

  1. When a user asks for the nearest store, the chatbot prompts them to share their location.

  2. Based on that input, the AI returns a list of nearby points of sale - sorted by proximity.

  3. Each result includes:

    • Store name

    • Opening hours

    • Phone number

Source: Chatarmin, Q3 2024

This way, customers can quickly find the most convenient location and get in touch or visit right away.

Make sure your chatbot responses are clear and actionable, so users can easily move from search to purchase!

BONUS 1: Watch a full Chatarmin AI-Chatbot-Built from Scratch, by our founder Johannes Mansbart, here:

Source: Chatarmin, Q3 2024

BONUS 2: Watch a full Chatarmin AI Setup Video by one of our gifted #teamarmin Devs, here:

Source: Chatarmin, Q2 2024

Optimal Training Methods for AI Systems

Summary

Step-by-Step Process

  1. Create Comprehensive Instructions

    • Formulate detailed guidelines and expectations

    • Define clear boundaries and desired response formats

  2. Data Collection and Integration

    • Systematically crawl websites

    • Upload all relevant manuals

    • Integrate customer inquiries and support tickets

    • Include document collections with typical queries

  3. System Integration

    • Link with existing ticket management systems

    • Implement API connections for real-time data exchange

  4. Continuous Improvement

    • Actively use the "Improve Answer" feature

    • Regularly review and evaluate responses

    • Pose simulated customer inquiries

    • Identify and close knowledge gaps

  5. Documentation and Monitoring

    • Log improvement processes

    • Capture performance metrics

    • Analyze and aggregate results

Through this iterative process, the AI continuously improves, learns from real interactions, and can provide increasingly precise and useful answers.

❗Troubleshooting Note: IF the AI does not react to customer messages, feel free to trigger the AI-flow manually out of chat.

Chatarmin AI Demo

Here you can test out our Chatarmin AI, which is used by our 7–8 figure clients to automate their WhatsApp sales, customer support, and more.
​
​MARC O'POLO
​https://waurl.me/tvt_b2ekr


​PHC Beauty
​https://waurl.me/nrf6dbwzx


​cusbclo
​https://waurl.me/pjkq6oqzm


​

Related Articles
Revenue attribution in Chatarmin
Chatbase - How to setup your Chatbot
Chatarmin DIXA Integration
WhatsApp Product Catalog
WhatsApp ads
Did this answer your question?