Manda4 Jarvis IA

The Manda4 Jarvis IA is an AI-powered conversational journey platform for Bitrix24: a visual (drag-and-drop) Flow Builder where each client builds their own service bot, no coding required. Combine AI and non-AI blocks in the same flow, with native, deep Bitrix24 integration.

Available building blocks

• Message — sends a text
  
• Menu — presents numbered options for the customer to choose from
  
• Question (free text) — collects a piece of data (name, ID number, email, date, phone, file...)
  
• Condition — branches the path based on a rule
  
• Bitrix24 Action — creates/updates a lead, deal, contact, calendar event
  
• External API call — integrates with any third-party system
  
• Handoff — routes to a queue or human agent
  
• End — closes out the service interaction
  
• A/B test — tests two versions of a step and measures the result
  
• Scheduled reminder — sends an automatic message after a set time (fixed or AI-generated)
  
• AI Agent — free-form conversation with the customer within a given topic (with or without a knowledge base)
  
• Knowledge base — documents/PDFs the AI consults to answer accurately
  
• Language detection — automatically identifies the customer's language
  
• Audio transcription — converts incoming audio into text
  

What the AI specifically does

• Understands free text, not just menu clicks (e.g., customer types "I want to pay" instead of pressing "1")
  
• Extracts data from unclear answers (e.g., separates the unit number from "apt 302, house 12")
  
• Answers open-ended questions within a topic, citing the knowledge base when configured
  
• Generates re-engagement messages on its own: if a deal stalls at a stage in the funnel, the AI reads the conversation history and sends a natural message to re-engage the customer
  
• Build copilot: you describe what you want in plain language and the AI assembles pieces of the flow for you to review before applying
  

WhatsApp/Bitrix24 connection
Each bot connects to a Bitrix24 Open Channel, respects a configurable business-hours schedule, and hands off to a queue/agent without the bot interfering again afterward.

Plans: