Messengers and Chatbots for Bitrix24

Open Channels: Where Everything Starts and Breaks

The "Open Channels" module (imopenlines) is Bitrix24's built-in mechanism for omnichannel communications. It works through the Im\Model\ChatTable entity, linking an external channel to an internal chat. The problem is that out-of-the-box routing settings are primitive: "round-robin" or "broadcast to all." For a real sales team with 15+ managers, VIP clients, and response time SLAs, that's not enough. We extend routing through OnImOpenLinesChatStart event handlers and REST API.

A manager switches between five windows, loses messages, forgets to reply — the client leaves for a competitor who responded in 30 seconds. Connecting messengers to Bitrix24 consolidates all channels into a single interface, while the CRM records every touchpoint.

Connecting Messengers

WhatsApp

The primary business channel. Integration via WhatsApp Business API — official connection with a verified account.

What works:

• Sending and receiving from the B24 interface — messages land in the open channel
• HSM templates for initiating conversations (abandoned cart reminders, order status). Templates go through Meta moderation — allow 2-3 days
• File, image, and document transfer
• Conversation linked to contact and deal via CRM_ENTITY_TYPE and CRM_ENTITY_ID

Connection options:

Telegram

Growing as a business channel. Especially in IT, education, and professional services. The Bot API is free and well-documented — a pleasant rarity.

What we do:

• Connect the bot to open channels via the imopenlines → "Telegram" connector
• Receive messages, photos, videos, documents — everything maps to a B24 chat
• Inline buttons and reply keyboards for bot navigation
• Webhook to https://yourdomain/rest/imconnector.register — connector registration
• CRM integration via REST API: incoming message → lead or deal activity

Where Telegram is irreplaceable:

• Support via bot — routine questions resolved without an operator
• Notifications: orders, delivery, payment — via Telegram Bot API sendMessage
• Internal team communications (though B24 chat exists for that)
• Lead generation: the bot asks qualifying questions → creates a lead via crm.lead.add

Viber

Maintains a strong position in certain regions and among the 35+ demographic.

• Business account connection via the open channels connector
• Viber Business Messages — bulk messaging with action buttons
• Rich content cards
• Sending and receiving from CRM

VK (VKontakte)

The largest social network in the CIS region. A mature toolset.

• Community messages connected via the imopenlines connector
• Message and comment processing from a unified interface
• Auto-creation of a lead from an inquiry — OnImOpenLinesCrmCreate handler
• VK Ads integration for source tracking via UTM
• Bot for auto-replies in the community via VK Bot API + Callback API

Open Channels Configuration: Details That Matter

All channels in one window: WhatsApp, Telegram, Viber, VK, live chat, Facebook Messenger, Instagram Direct, Email, Avito, Yandex.Market (via third-party connectors).

Settings that are often overlooked:

• Distribution: round-robin, by workload (QUEUE_TYPE), by competency — via a custom handler
• Auto-greeting — configured in the channel's "Automatic Actions" section
• Conversation transfer between operators — the client doesn't see the switch
• Post-conversation quality rating (CSAT) — widget in the chat
• Business hours and after-hours auto-reply — linked to the schedule from the HR module
• Wait timeout: if an operator doesn't pick up the chat within N minutes — escalation to the next in queue

Routing

Who responds to a specific inquiry? By default — "whoever is free." In reality, it needs to be more sophisticated:

• Identify the responsible manager by phone/email from CRM — im.chat.get + lookup via crm.contact.list
• Department-based distribution using keywords (NLP classifier or simple regex on the first message)
• Priority queue for VIPs — by CRM segment
• Escalation on 5-minute timeout — automatic transfer to the next agent
• Switch to a call directly from the chat — telephony.externalcall.register

Chatbots

Types

Rule-based (scenario-driven): Button menu, decision tree. "How to pay" → "Where's my order" → "Business hours." Handoff to an operator when the bot can't handle it — by condition if intent == 'unknown' → transfer_to_queue. Reliable, predictable, covers 60-70% of routine inquiries.

NLP-powered: Free-text input. Intent detection (buy, complain, ask about delivery), entity extraction (name, date, order number). Contextual dialogue — remembers what was discussed. Built on Rasa or Dialogflow, integrated with B24 via REST.

Scenarios and Real Impact

Bot Development

1. Inquiry analysis — we export history from open channels and cluster by topic. We find what consumes 80% of operators' time.
2. Dialogue design — flowchart in Miro/Figma. The bot sounds like a person, not a microwave manual. Every branch ends with either an answer or a handoff to an operator.
3. Development — logic, CRM integration, and external APIs. For rule-based bots — a finite state machine. For NLP — pipeline: tokenizer → featurizer → classifier → response selector.
4. NLP training — training data from real conversations, intent recognition tuning, confidence threshold for fallback.
5. Testing — running through all branches, edge cases (empty message, sticker, voice message).
6. Optimization — monitoring fallback rate, retraining on new conversations.

Auto-Replies and Templates

• Greeting on first contact — the client immediately knows they've been heard
• After-hours reply with business hours
• Receipt confirmation with estimated SLA
• Keyword-triggered auto-reply — configured in the channel's "Automatic Actions"
• Template library for operators — hotkeys, name and order number substitution from CRM via #CONTACT_NAME#, #ORDER_ID#

CRM Integration

• First contact → lead (crm.lead.add) or contact. Duplicates checked by phone/email via crm.duplicate.findbycomm
• Repeat contact → linked to existing contact, activity in timeline
• Auto-creation of a deal by trigger (specific keyword or dialogue stage)
• Entire conversation history — in the client card, chronologically
• CRM marketing: segmented campaigns, trigger chains (abandoned cart, welcome series), personalization from CRM fields

NLP and AI

• Inquiry classification: tech support / sales / return — automatically
• Long conversation summarization for managers — via GPT API with a prompt for extracting the key points
• Context-aware suggestions for the operator
• Sentiment analysis: satisfied / frustrated / neutral — affects queue priority
• Auto-translation for multilingual support

Analytics

• Average first response time (FRT) by channel — department KPI
• Average resolution time (ART)
• Volume by channel and topic — understanding where to grow
• CSAT — satisfaction score
• Operator workload — data-driven staffing decisions

Timeline

Result — all inquiries in a single window, routine tasks automated, no message gets lost. Managers sell instead of hunting for the right chat.