Uniflow: The Visual Workflow Engine of Uniconnect

Uniflow is the dynamic automation engine behind Uniconnect, enabling teams to visually build, trigger, and manage powerful business workflows across the ecosystem — from support and sales to engagement and AI-powered insights.

---

Core Concepts

Trigger-Driven or Scheduled Workflows

Every Uniflow starts with either:

• a trigger — an event that initiates the workflow, or
• a cron schedule — a recurring timer (e.g., every day at 9 AM)

These triggers can originate from any Uniconnect module:

Sample Triggers:

• Call started / ended
• Email received
• Ticket created / updated
• User logged in
• SMS received
• Campaign started / stopped
• Activity status changed

Sample Triggers:

• Call started / ended
• Email received
• Ticket created / updated
• User logged in
• SMS received
• Campaign started / stopped
• Activity status changed

---

Node Types

Uniflow supports multiple types of nodes, each defining a different kind of logic:

Node Type	Description
trigger	Starts the workflow
action	Performs an operation (e.g., send email)
condition	Branches based on logic (if statements)
delay	Waits for a specified duration
loop	Iterates over a list
parallel	Executes multiple paths simultaneously

---

Example Actions

Actions in Uniflow interact with the Uniconnect ecosystem:

• email.send: Send an email
• sms.send: Send an SMS
• ticket.create: Create a support ticket
• campaign.addUser: Add user to a campaign
• campaign.start: Start a campaign
• activity.changeStatus: Update a record status

---

AI-Augmented Nodes

Uniflow integrates with AI services to unlock intelligent, context-aware decision-making and automation. These AI nodes can be powered by OpenAI, HuggingFace models, internal LLMs, or custom inference APIs.

Available AI Nodes

• ai.extract
  Convert unstructured input (e.g., emails, call transcripts, form text) into structured JSON data for use in workflows.

• ai.classify
  Categorize content (e.g., route tickets by topic, detect sentiment, identify lead type, prioritize messages).

• ai.sentiment
  Analyze emotional tone (positive, neutral, negative). Useful for prioritizing angry customers, detecting churn risk, or personalizing replies.

• ai.decide
  Make decisions based on context (e.g., should we escalate? is this a new issue or follow-up? should we retry or abort?).

• ai.generate
  Create natural-language responses (e.g., email drafts, chatbot replies, SMS follow-ups, task summaries).

• ai.match
  Find the best match from a list (e.g., assign to most relevant agent, map to best-fit knowledge base article, recommend a product).

• ai.search
  Semantic search over documents, FAQs, call notes, etc. (e.g., retrieve useful context for ticket resolution).

• ai.summarize
  Condense long conversations, call transcripts, or ticket history into a quick summary for agents or reporting.

• ai.enrich
  Add context to minimal inputs (e.g., enrich a lead with company info, or complete a partial form).

• ai.validate
  Check for inconsistencies, completeness, or correctness of submitted data or messages.

Example AI Workflows

• Email Routing
  Use ai.classify to determine the department, ai.extract to pull the ticket data, and ai.decide to assign priority.

• Bot Response + Agent Handoff
  Use ai.generate for initial responses and ai.decide to escalate when needed.

• Smart Summary Before Escalation
  Before escalating, use ai.summarize on ticket history and call notes, then send it to the supervisor.

• Context-Aware Reply
  Use ai.search to find recent similar cases, then ai.generate to compose a personalized reply.

---

Extensibility

Uniflow is modular. You can:

• Add custom triggers and actions from internal services
• Integrate third-party APIs (e.g., CRMs, call centers)
• Expand AI capabilities using prompt templates or external ML models

---

Summary

Uniflow is the automation backbone of Uniconnect — combining no-code visual building, deep ecosystem integration, and powerful AI to enable smart, efficient, and personalized experiences.