AI Agents

Dataiku provides a full environment for building, troubleshooting, evaluating, deploying, monitoring, and exposing at scale AI Agents for the enterprise.

• Introduction to Agents in Dataiku
  • Tools
  • Types of agents
    • Visual Agents
    • Code Agents
    • Other agent kinds
  • Using agents
• Simple Visual Agents
  • Sources
  • Context
• Code Agents
  • Requirements
  • What does the Code do
  • Using tools
  • Streaming
• Structured Visual Agents
  • Overview and example use cases
    • Predefined sequence of steps and branches
    • Gathering “mandatory” information at the beginning of the conversation
  • Key concepts
    • Blocks
    • State and Scratchpad
    • Expressions and templating
    • Starting block and next turn behavior
    • Pre-turn and post-turn blocks
  • Blocks
    • Core Loop
      • Exit Conditions
        • Example use case
      • State and scratchpad virtual tools
      • Forcing tool call arguments
      • Using Agents as tools
      • History passing
    • Routing
    • Manual and Mandatory Tool Call
    • Parallel
      • Output
      • Cautions
    • For Each
      • Example use case
      • Output
      • Cautions
    • Set State and Scratchpad entries
    • Emit Output
    • Long-Term Memory
      • Setup
      • Types of memory items
      • Memory banks
    • Semantic Feedback Management
      • Setup
      • Usage
  • Expressions and templates
    • CEL expressions
    • CEL templates
    • Jinja
  • How to …
    • Conversational Disambiguation: Force an agent to uniquely identify the subject of discussion before moving on
• Managed tools
  • Using tools
    • Visual Agent
    • API
      • Native API
      • LangChain API
      • Tool descriptor
      • Context
  • Knowledge Bank Search tool
    • Core configuration
    • Retrieval settings
    • Sources
    • Filtering
      • Static filtering
      • Dynamic filtering
      • Agent-inferred filtering
    • Document-Level Security
  • Semantic Model Query Tool
  • Dataset Lookup tool
    • Configuration
    • Usage
    • Limitations
  • Dataset Append tool
  • Local MCP
    • Prerequisites
    • Configuration
    • Usage
    • Security
  • Remote MCP
    • Prerequisites
      • Authentication
    • Configuration
    • Usage
  • Model Predict tool
  • API Endpoint Tool
  • Query an LLM/Agent
  • Send Message tool
    • Message variables
  • Calculator tool
  • SQL Question Answering Tool
    • Overview
    • Getting Access
    • Configuration
      • Introduction
      • Requirements
        • Infrastructure
    • Tool Instance: Design
      • Basic Settings
      • Dataset Information & Descriptions
      • Return Values
      • Sample Values
      • Tool Descriptor Settings
  • Google Search Tool
  • Google Calendar
    • Create Google Calendar Event
  • Optimization Tool
    • Setup
    • Usage
    • Inputs
    • How It Works
    • Output
  • Jira
    • Create Jira Issue
  • Salesforce
    • Lookup Salesforce Account
    • Create Salesforce Contact
  • ServiceNow
    • Create ServiceNow Issue
  • Snowflake Cortex Analyst Tool
    • Prerequisites
    • Usage
  • Snowflake Cortex Search Tool
    • Prerequisites
  • Databricks Genie Tool
    • Prerequisites
    • Usage
  • Databricks Vector Search Tool
    • Prerequisites
    • Usage
  • Writing your own tool
    • Inline code tool
    • Plugin tools
• External Agents
  • Snowflake Cortex Agents
    • Prerequisites
    • Usage
  • Databricks Agents
    • Prerequisites
    • Usage
    • Limitations
  • AWS Bedrock Agents
    • Key Features
    • Prerequisites
      • AWS Requirements
    • Usage (Bedrock Agent)
    • Usage (Bedrock AgentCore)
    • Cautions
  • Google Vertex AI Agents
    • Prerequisites
    • Usage
    • Cautions
• Agent Hub
  • Overview
  • Getting Access
  • Configuration
    • Requirements
    • Initial Setup
    • Backend Settings
    • Webapp Settings
      • Core Settings
      • Enterprise Agents
      • My Agents
      • Prompt Library
      • Charts Generation
      • Upload in Conversation
      • User Experience
      • Look and feel customization
    • Setting up integrations with Stories and Traces Explorer
      • Traces explorer
      • Stories
    • Building tools that integrate with Agent Hub
      • Display Sources
      • Generate graphs
      • Downloadable Files
  • Agent Hub Ops
  • Using Agent Hub
    • Agent Library
    • Conversations with Agents
    • Building “My Agents”
    • Understanding Responses
  • Monitoring and Quality checks
    • Extracting Agent Hub’s logs
    • Monitoring
    • Quality checks
  • Security and Permissions
    • Webapp permissions
    • Enterprise agents permissions
    • Document-level security
    • Role-based access control
    • User Access Profiles
• Agent Review
  • Overview
  • Tests
    • Adding tests
    • Quick Test
    • Running tests
  • Traits
  • Review run results
    • Analyze results
    • Test Annotation & Feedback
    • Test status
    • Trait status
  • Compare runs
  • Settings
    • Multiple execution of a test
    • Trait setup
    • Permissions
• Tracing
  • Anatomy of a trace
  • Traces built into the LLM Mesh
  • Adding your own trace items
  • Trace Explorer
    • Creating a Trace Explorer Webapp
      • Use with Prompt Recipe, Agent Connect and Dataiku Answers
    • Using the Trace Explorer
    • Accessing via Explore Trace shortcut
    • Troubleshooting & Notes
• Agent Evaluation
  • Overview
  • Recipe configuration
    • Input dataset
    • Outputs
    • Metrics
    • Custom Metrics
    • Custom Traits
  • Agent Evaluations
  • Trajectory explorer
  • Comparisons
• Additional Request Context
  • Standard Keys
  • Usage
• Slack Integration
  • Overview
  • Setup
    • Pre-requisites
    • Create a Slack App
      • Option 1: Quick Start (Manifest)
      • Option 2: Standard Setup (From Scratch)
    • Generate an App Level Token
    • Install the Slack App to your Slack Workspace
    • Create and Configure the Visual Webapp in Dataiku
    • Test the Integration
    • Adding Bots to Channels