AI Agents

Connect AI tools and external agents to your Precog data platform.

Precog exposes machine-friendly protocols so that AI assistants, automation workflows, and other agents can explore your data infrastructure, monitor freshness, troubleshoot issues, and recommend datasets — all without logging into the Precog UI.

Available Integrations

Setup Wizard
Create an AI Assistant in minutes
Walk through the guided wizard to connect your data, describe your business context, and configure your assistant — all in one flow.
Chat Setup
Connect your AI tools
Get your MCP server URL and per-client setup instructions for Claude, ChatGPT, GitHub Copilot, Cursor, Gemini CLI, and more.
MCP Server
Model Context Protocol for AI assistants
Connect Claude Desktop, Cursor, GitHub Copilot, Gemini CLI, ChatGPT, Microsoft Copilot Studio, and other MCP-compatible tools directly to your Precog workspace.
Knowledge Base
Curate your business knowledge
Review, approve, and manage the business facts (KPIs, definitions, rules) that teach your AI Assistant about your domain.
Semantic Layer
Explore your semantic model
Visualize datasets, fields, relationships, and metrics. Track generation status and regenerate when your knowledge base changes.
Sessions & Monitoring
Track assistant activity
Monitor session counts, tool call volumes, and per-user activity to understand adoption and spot issues.
Agent Access and Security
Control warehouse identity and access
Add warehouse identities, choose a default, and assign a principal to each member who uses an AI agent.

Getting Started

The fastest way to get up and running is to use the Setup Wizard — it walks you through connecting your data, describing your business context, and provisioning your assistant in a single flow.

All integrations use OAuth 2.1 for authentication. Interactive AI clients (like Claude Desktop, Cursor, and GitHub Copilot) handle the OAuth flow automatically — you just provide the MCP server URL and sign in when prompted. For automated or server-to-server access, create a Machine Client in the admin panel.

  1. Create an AI Assistant using the setup wizard
  2. Connect your AI tools using the Chat Setup page
  3. Start asking questions about your data

Capabilities

When connected through any protocol, AI tools can:

CapabilityDescription
Data ExplorationList workspaces, sources, destinations, connections, and datasets
Freshness MonitoringCheck data freshness, detect stale data, and review sync errors
Connection TroubleshootingInvestigate run history, diagnose failures, and understand sync timing
Dataset RecommendationsDiscover available datasets and get suggestions for what to sync
Context ManagementRecord and retrieve organization-specific terminology and business rules
Team OverviewView team members and their roles