NEWSRelevance Lab makes a strategic majority investment in kaman.aiLearn more

kaman.ai

IntegrationsPricingTemplatesBlogPapersDocs

Documentation

Guides, use cases & API reference

  • Overview
    • Getting Started
    • Platform Overview
  • Features
    • Features Overview
    • AI Assistant
    • Workflow Automation
    • Intelligent Memory
    • Data Management
    • Universal Integrations
    • Communication Channels
    • Collective Agent Memory (CAML)
    • Security & Control
  • Use Cases Overview
  • Financial Services
  • Fraud Detection
  • Supply Chain
  • Technical Support
  • Software Development
  • Smart ETL
  • Data Governance
  • ESG Reporting
  • TAC Management
  • Reference
    • API Reference
  • Guides
    • Getting Started
    • Authentication
  • Endpoints
    • Workflows API
    • Tools API
    • KDL (Data Lake) API
    • OpenAI-Compatible API
    • A2A Protocol
    • Skills API
    • Knowledge Base (RAG) API
    • Communication Channels
    • DSL Import API
Release Notes

Workflows

Workflows are multi-step automations that chain together functions, conditions, and data transformations. Use the visual editor to build directed acyclic graphs (DAGs) that define how data flows through your automation.


Building a Workflow

Click the + Create button to open the workflow editor:

  1. Add nodes — drag functions, conditions, or LLM steps onto the canvas
  2. Connect edges — draw connections between nodes to define data flow
  3. Configure inputs — set each node's input parameters and map outputs from previous steps
  4. Test and run — execute the workflow and monitor results

Node Types

Node TypePurposeExample
FunctionExecute code or call an APIFetch customer data, send email
ConditionBranch logic based on a valueIf status == "approved", go to step A
LLMProcess text with an AI modelSummarize a document, classify intent
TransformReshape or filter dataMap fields, merge datasets

Triggering Workflows

Workflows can be started in several ways:

  • Manual — run from the UI or via chat
  • API — trigger programmatically via REST endpoint
  • Schedule (Cron) — run on a recurring schedule
  • Webhook — trigger from external events
  • Agent — the agent can invoke a workflow as part of a conversation

Credential Mapping

Workflows that use tools requiring authentication need credential profiles mapped to each tool:

json
{
  "tool_123": "analytics_readonly",
  "mcp_456": "postgres_admin"
}

This ensures each step uses the right credentials without exposing secrets in the workflow definition.

Monitoring Runs

Each workflow execution creates a run log with:

  • Status — success, failed, or running
  • Step-by-step results — what each node produced
  • Error details — where failures occurred and why
  • Timing — how long each step took

Quick Tips

  • Start simple — build and test one step at a time
  • Use the visual editor to understand data flow between nodes
  • Check workflow run logs to debug failed executions
  • Map credential profiles to tools that need authentication
  • Use condition nodes to handle edge cases and errors gracefully
  • Export proven workflows as templates for reuse

On this page

  • Building a Workflow
  • Node Types
  • Triggering Workflows
  • Credential Mapping
  • Monitoring Runs
  • Quick Tips