Skip to content
Rush Commerce
Software & Dev2 min read

Boomi buys Elementary Data: data observability for AI agents

Boomi acquired Elementary Data, the open-source data observability tool with an MCP server for AI agents. Why bad data, not the model, breaks your agents.

Integration platform Boomi acquired Elementary Data, an Israeli data observability company that started as an open-source project. Boomi's pitch is blunt: AI agents are only as good as the data under them. Elementary watches that data and catches bad rows before they reach a dashboard, a model or an agent. When an integration vendor pays real money for data quality, the market is telling you where agents break.

What actually happened

Per Boomi's announcement on October 19:

  • The product: Observability, quality checks, governance and discovery. It includes column-level lineage, a data catalog, health scores, and an MCP server that exposes data health to AI tools and agents.
  • The open-source base: Elementary launched in 2022 as an open-source project. Boomi says it's used by 1,500+ data teams with 1.1M+ monthly downloads. Elementary Cloud added paid features in 2024.
  • Price: Boomi didn't disclose terms. The Philadelphia Inquirer, citing Israel's Calcalist, put it at more than $40 million. The Inquirer says Elementary has 17 employees and co-founder Maayan Salom joins Boomi with the team.
  • Boomi's own data: A Forrester study Boomi commissioned (July 2026) found 86% of leaders say they're past agent pilots, but only 34% trust what their agents do.

Boomi's announcement doesn't spell out the open-source project's long-term future. If you run it, watch the repo and the license.

Why data observability matters for your business

Most agent failures we see aren't model failures. The agent read a stale price, a duplicate customer or a null SKU and did exactly what it was told with bad facts. A better model doesn't fix that. Better inputs do.

You don't need Boomi's platform to act on this:

  1. Test the tables your agent reads. Freshness, nulls, duplicates, row counts. If you use dbt, Elementary's open-source package does this today.
  2. Make the agent check health first. An agent that reprices products should refuse to run when the inventory sync is six hours late. That's one query before the action.
  3. Know your exit. Acquired open-source tools sometimes change license or slow down. Pin versions and keep your tests in plain SQL you own.

Key takeaways

  • Boomi acquired Elementary Data, an open-source data observability tool with an MCP server for agents
  • Elementary reports 1,500+ data teams and 1.1M+ monthly downloads
  • Price undisclosed by Boomi; reported at more than $40 million
  • Boomi-commissioned Forrester research: 86% past pilots, 34% trust their agents' actions
  • Test the data your agents read, and make them check freshness before they act

Agent making odd calls? Check the data before you blame the model. We wire freshness and quality checks into the agents we build, in SQL you own, not a vendor's black box. See how we build, or send us the agent that's misbehaving.

Sources: Boomi, The Philadelphia Inquirer.

  • #data-observability
  • #ai-agents
  • #boomi
  • #mcp
  • #data-quality
TR

Tommy Rush — Founder, Rush Commerce

Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More

Get The Rush Report weekly — one email, zero fluff.