Portal Community
What it does: The Chat node suspends workflow execution, delivers a message to one or more recipients through a configured communication channel, and waits for a text reply. When the recipient replies, the built-in IChatReplyParser AI classifier analyses the free-text response and categorises it as approved, rejected, or unknown. The workflow resumes on the matching output port, carrying the original reply text and classification result.
AI-Powered Classification: Unlike the Approval node which requires a structured inbox interaction, the Chat node accepts any natural language reply. The IChatReplyParser uses a language model to determine whether the reply represents a positive (approved) or negative (rejected) intent. For example, "Sounds good, go ahead!" classifies as approved, while "No, hold off on that" classifies as rejected. Unclassifiable replies route through the success port for custom handling.

Supported Channels

The Chat node delivers messages through your organisation's integrated communication platforms:

Slack WhatsApp Email SMS Teams

Channel availability depends on which integrations are configured in your BizFirstAI environment. If the preferred channel is unavailable for the recipient, the node will attempt fallback to email.

Two-Phase Execution

Like all HIL nodes, the Chat node operates in two phases separated by an open-ended suspension. The workflow is completely dormant between phases — it does not poll, does not consume resources, and survives server restarts.

Phase 1 — Send & Suspend

Message Delivery

  • Resolves actor identity and preferred channel
  • Delivers the message via the configured channel integration
  • Registers a reply listener (webhook or polling) for incoming messages
  • Suspends workflow; "waiting" port fires
  • Timeout clock starts
Phase 2 — Receive & Classify

AI Classification

  • Incoming reply detected by the channel listener
  • IChatReplyParser analyses the reply text
  • Classification result: approved / rejected / unknown
  • Workflow resumes on matching port
  • Reply text and classification are available in node output

Key Capabilities

Business Benefits

Many approval and confirmation scenarios do not warrant the friction of a formal approval portal. When a manager needs to quickly confirm a deployment, a field technician needs to report job completion, or a customer needs to confirm their order details, a simple chat message is far more accessible and actionable than a formal approval inbox.

The Chat node bridges this gap by meeting users where they already are — in Slack, WhatsApp, or their email client. The AI classifier removes the need for structured responses ("Reply YES or NO"), allowing natural communication. This dramatically improves response rates compared to form-based approvals, particularly for time-sensitive or mobile-first use cases.

For customer-facing workflows, the Chat node enables WhatsApp-based confirmations that are embedded directly in the automated process. A customer confirms their delivery time, validates their order, or accepts a quote without needing to log into a portal — their reply in WhatsApp is sufficient to drive the next step of the business workflow.

Use Cases

Quick Deployment Confirmation via Slack

Before deploying to production, send a Slack message to the on-call DevOps engineer: "Ready to deploy version 2.4.1 to production. Reply YES to proceed or NO to cancel." Their reply is classified and the workflow either triggers the CI/CD pipeline or halts and notifies the release manager.

Customer Order Confirmation via WhatsApp

After an order is placed, send a WhatsApp confirmation to the customer: "Hi {{name}}, your order #{{order_id}} for ${{total}} is ready to ship. Reply CONFIRM to proceed or CANCEL to stop. The customer's natural language reply is classified and triggers either shipping or order cancellation workflows.

Manager Quick Approval via Email Reply

For non-urgent approvals, email the manager a summary of the request and ask them to reply to the email. "Reply APPROVE or REJECT with any comments." The email reply is captured, classified by the AI, and the workflow continues accordingly — the manager never needs to leave their inbox.

Field Agent Status Update via SMS

After assigning a field service ticket, send the technician an SMS with job details: "Job #{{ticket_id}} assigned: fix router at {{address}}. Reply ACCEPT to confirm or UNAVAILABLE to reassign." The reply is classified and routes the ticket to the confirmed technician or triggers reassignment logic.

Executive Sign-Off via Teams Message

For board-level decisions that arise during off-hours, send a Teams message to the executive: "Urgent: vendor contract with {{vendor}} expires at midnight. Reply EXTEND to renew for 30 days or ESCALATE to initiate formal review." Extends or escalates based on the classified reply, with a full audit record.

In This Guide

Configuration

Actors, channel selection, message templates, timeout settings, and expression support reference.

Input & Output

Output port reference, reply data schema, classification values, and downstream access patterns.

Examples

Complete configurations for Slack deployments, WhatsApp customer flows, email approvals, and SMS field dispatch.