The term “AI-native” gets thrown around a lot. Usually it means “we added an LLM endpoint.” That’s not what we mean by agent-native.

An agent-native data platform is one where AI agents are first-class operators — not bolted-on consumers of a human-designed API, but autonomous participants who can discover, configure, ingest, validate, transform, and query data through the same interfaces humans use.

The Problem with Bolt-On AI

Most data platforms today treat AI as a feature. You get a chatbot that can answer questions about your pipeline status, or an LLM that generates SQL. That’s useful, but it’s fundamentally limited — the AI is a passenger, not a driver.

When an AI agent needs to:

  • Discover what datasets exist and their schemas
  • Create a new ingestion pipeline from scratch
  • Set up validation rules in plain English
  • Route data to multiple destinations
  • Diagnose why a pipeline failed at 3 AM

…it shouldn’t need a human to write glue code, configure webhooks, or build a custom integration. It should just talk to the platform.

What Agent-Native Looks Like

At Datris, we built Model Context Protocol (MCP) support directly into the platform. MCP is becoming the standard for how AI agents interact with tools and data sources — think of it as USB-C for AI.

Through MCP, an agent can:

  • List all datasets and inspect their configurations
  • Create new datasets with AI-generated schemas
  • Upload and ingest files — CSV, JSON, XML, Excel, PDFs
  • Set validation rules in plain English — “reject rows where email format is invalid”
  • Run AI-powered data profiling — get summary stats, quality issues, and suggested rules
  • Query processing history and diagnose errors
  • Route data to PostgreSQL, MongoDB, vector databases, message queues — all configured through conversation

No integration code. No middleware. The agent talks to the platform the same way a developer would use the API — but through natural language.

Why This Matters Now

Three trends are converging:

  1. Agents are getting autonomous. Claude, GPT, and open-source models can now plan multi-step workflows, use tools, and recover from errors. They’re ready to operate infrastructure.

  2. Data pipelines are getting more complex. The average enterprise runs dozens of ingestion jobs across multiple sources, formats, and destinations. Managing this manually doesn’t scale.

  3. MCP is creating a standard. Before MCP, every AI integration was bespoke. Now there’s a protocol that lets any agent talk to any tool. Platforms that support it become instantly accessible to the entire AI ecosystem.

The First, Not the Last

We believe Datris is the first open-source data platform with native MCP support. We won’t be the last — and that’s the point. The industry is moving toward agent-operated infrastructure. The platforms that embrace this early will define the category.

If you’re building data pipelines and thinking about how AI agents fit into your architecture, check out the repo or reach out. We’d love to compare notes.

Todd Fearn is the founder of Datris.ai and has been building AI solutions and data infrastructure for financial services for 25+ years, including at Goldman Sachs, Bridgewater Associates, Deutsche Bank, Freddie Mac and others.