Skip to content

PATTERN Cited by 1 source

Dual interface — human and agent

Pattern

Expose the same governed data and context layer through two interfaces — a rich UI for human discovery/interaction and an MCP (or equivalent machine-readable) server for AI agents — backed by the same identity, governance, and authorization system.

Structure

┌─────────────┐     ┌───────────────────────┐     ┌─────────────────┐
│  Human user │────▶│  Marketplace / Web UI │────▶│                 │
└─────────────┘     └───────────────────────┘     │  Governance     │
                                                   │  (Unity Catalog)│
┌─────────────┐     ┌───────────────────────┐     │                 │
│  AI Agent   │────▶│  MCP Server           │────▶│                 │
└─────────────┘     └───────────────────────┘     └─────────────────┘

Key property: the agent sees exactly what the human sees (and nothing more). There is no second, looser access path around the governance system.

Why this matters

  • Governance parity — one access model means one audit trail, one policy surface, one identity system
  • No shadow paths — agents don't accumulate permissions humans lack
  • Trust — engineers can trust agent answers because they know the agent operated within the same boundary

"Employees can discover data products, see owners and lifecycle status, open linked dashboards and apps, query Unity Catalog-governed data, and use a chat interface for natural-language exploration. For AI clients, the same context is exposed through MCP." (Source: sources/2026-07-21-databricks-why-rd-data-belongs-in-the-lakehouse-and-why-agents-need-it)

Seen in

Last updated · 590 distilled / 1,788 read