AI Driven Automation in
Open-Source Metadata Platforms
Embedding an MCP Server
● Introducing OpenMetadata
● AI Agents need for Context & MCP Servers
● AI-driven governance with OpenMetadata MCP Server
● Workshop Pt 1
● Coffee Break
● Workshop Pt 2
● Q&A
Agenda
prereqs
bit.ly/omd-odsc
Data complexity is growing
INGEST TRANSFORM
RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS
DATA SOURCES
DATA
INFRASTRUCTURE
Unified Metadata Graph Powered by
All your metadata: collected and standardized
All your metadata: collected and standardized
Unified Metadata Graph Powered by
Discovery Observability Governance
Discovery Observability
Unified Metadata Graph Powered by
Open Source
API-First and
Schema-First
Open Standards
Metadata
Extensibility
Governance
All your metadata: collected and standardized
INGEST TRANSFORM
RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS
DATA SOURCES
DATA
INFRASTRUCTURE
Discovery Observability
Unified Metadata Graph Powered by
Governance
All your metadata: collected and standardized
SOURCE
LINEAGE
INCIDENT
LINEAGE
DESCRIPTION
QUERIES
SAMPLE
GLOSSARY
PIPELINES
INGEST TRANSFORM
RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS
DATA SOURCES
DATA
INFRASTRUCTURE
DATA
ANALYSTS
DATA
ENGINEERS
BUSINESS
USERS
DATA
SCIENTISTS
DATA
GOVERNANCE
Data contextualized for every user and every agent
DATA
ANALYSTS
DATA
ENGINEERS
BUSINESS
USERS
DATA
SCIENTISTS
DATA
GOVERNANCE
Data contextualized for every user and every agent
AI
AGENTS
DATA
ANALYSTS
DATA
ENGINEERS
BUSINESS
USERS
DATA
SCIENTISTS
DATA
GOVERNANCE
Data contextualized for every user and every agent
AI
AGENTS
Workflows for everyone
deliver high quality outcomes efficiently
Breakdown data silos
comprehensive context of data
100+ integrations
with leading data platforms & tools
Unified Metadata Graph Powered by
Workflows for everyone
deliver high quality outcomes
efficiently
Breakdown data silos
comprehensive context of data
100+ integrations
with leading data platforms & tools
DATA
ANALYSTS
DATA
ENGINEERS
BUSINESS
USERS
DATA
SCIENTISTS
DATA
GOVERNANCE
Discovery Observability Governance
Unified Experience
Data contextualized for every user and every agent
AI
AGENTS
INGEST TRANSFORM
RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS
DATA SOURCES
DATA
INFRASTRUCTURE
CONFIDENTIAL - DO NOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE
Loggi transforms data for more efficient
and reliable deliveries with
CONFIDENTIAL - DO NOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE
30% faster ETL pipelines
89% fewer Looker dashboards
7,000 Redshift tables dropped
$2,000 savings per month
with
CONFIDENTIAL - DO NOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE
Loggi transforms data for more efficient
and reliable deliveries with
30%faster ETL pipeline
89%fewer Looker dashboards
7,000Redshift tables dropped
$2,000savings per month
OpenMetadata has been useful since day one in the
company. Everyone claimed it was difficult to find
tables for specific information needed, or even to
know who they could reach out to in order to get
more information. Now, all our data components are
in one place and our users have been giving good
feedback about the experience.
The Power of Open Standards and Open Source
11,000 2,000 7,800
Community
members
Enterprise
deployments
GitHub
stars
370 100
Open-source
contributors
Data
connectors
{ }
https://slack.open-metadata.org/ https://github.com/open-metadata/OpenMetadata
Give agents context to
make the right
decisions
Standardizes the
agent<>tool interface
Conversational
interface for every
technical tool
Model Context Protocol - MCP
Out-of-the-box
deployment
OpenMetadata MCP Server
Out-of-the-box
deployment
Exposing functional
tools not an API
wrapper
OpenMetadata MCP Server
Out-of-the-box
deployment
Exposing functional
tools not an API
wrapper
Native
authentication &
authorization
OpenMetadata MCP Server
Automate changes up
and down a data stack
Combining with other
MCP servers
OpenMetadata MCP Server Use Cases
Murat Migdisoglu, Data Architect &
contributor
Aimen Denche, Data Engineer @
workshop
bit.ly/omd-odsc
Announcement
goose
Recipes
Learning Outcomes
OpenMetadata is an
open-source metadata
platform provides
discoverability,
observability, and
governance for every
data asset in your
organization to every
one and every agent
Model Context Protocol
standardizes the
interface between
agents and tools, giving
AI the right context to
make the right
decisions
OpenMetadata MCP
Server brings the
context and
governance of your full
data stack to AI agents,
allowing them to take
actions the improve
your data systems at
scale
Q&A
Star us on GitHub
https://github.com/open-metadata/OpenMetadata
Join our Slack
https://slack.open-metadata.org/
Follow us on X
@open_metadata
Collate is OpenMetadata aaS
https://www.getcollate.io/
Thank you

ODSC West 2025 OpenMetadata Workshop.pdf

  • 1.
    AI Driven Automationin Open-Source Metadata Platforms Embedding an MCP Server
  • 2.
    ● Introducing OpenMetadata ●AI Agents need for Context & MCP Servers ● AI-driven governance with OpenMetadata MCP Server ● Workshop Pt 1 ● Coffee Break ● Workshop Pt 2 ● Q&A Agenda
  • 3.
  • 4.
    Data complexity isgrowing INGEST TRANSFORM RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS DATA SOURCES DATA INFRASTRUCTURE
  • 5.
    Unified Metadata GraphPowered by All your metadata: collected and standardized
  • 6.
    All your metadata:collected and standardized Unified Metadata Graph Powered by Discovery Observability Governance
  • 7.
    Discovery Observability Unified MetadataGraph Powered by Open Source API-First and Schema-First Open Standards Metadata Extensibility Governance All your metadata: collected and standardized INGEST TRANSFORM RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS DATA SOURCES DATA INFRASTRUCTURE
  • 8.
    Discovery Observability Unified MetadataGraph Powered by Governance All your metadata: collected and standardized SOURCE LINEAGE INCIDENT LINEAGE DESCRIPTION QUERIES SAMPLE GLOSSARY PIPELINES INGEST TRANSFORM RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS DATA SOURCES DATA INFRASTRUCTURE
  • 9.
  • 10.
  • 11.
    DATA ANALYSTS DATA ENGINEERS BUSINESS USERS DATA SCIENTISTS DATA GOVERNANCE Data contextualized forevery user and every agent AI AGENTS Workflows for everyone deliver high quality outcomes efficiently Breakdown data silos comprehensive context of data 100+ integrations with leading data platforms & tools
  • 12.
    Unified Metadata GraphPowered by Workflows for everyone deliver high quality outcomes efficiently Breakdown data silos comprehensive context of data 100+ integrations with leading data platforms & tools DATA ANALYSTS DATA ENGINEERS BUSINESS USERS DATA SCIENTISTS DATA GOVERNANCE Discovery Observability Governance Unified Experience Data contextualized for every user and every agent AI AGENTS INGEST TRANSFORM RAW DATA REFINED / FEATURE DATA ANALYTICS / MODELS DATA / AI APPS DATA SOURCES DATA INFRASTRUCTURE
  • 13.
    CONFIDENTIAL - DONOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE Loggi transforms data for more efficient and reliable deliveries with
  • 14.
    CONFIDENTIAL - DONOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE 30% faster ETL pipelines 89% fewer Looker dashboards 7,000 Redshift tables dropped $2,000 savings per month with
  • 15.
    CONFIDENTIAL - DONOT DUPLICATE OR DISTRIBUTE WITHOUT PERMISSION FROM COLLATE Loggi transforms data for more efficient and reliable deliveries with 30%faster ETL pipeline 89%fewer Looker dashboards 7,000Redshift tables dropped $2,000savings per month OpenMetadata has been useful since day one in the company. Everyone claimed it was difficult to find tables for specific information needed, or even to know who they could reach out to in order to get more information. Now, all our data components are in one place and our users have been giving good feedback about the experience.
  • 16.
    The Power ofOpen Standards and Open Source 11,000 2,000 7,800 Community members Enterprise deployments GitHub stars 370 100 Open-source contributors Data connectors { } https://slack.open-metadata.org/ https://github.com/open-metadata/OpenMetadata
  • 17.
    Give agents contextto make the right decisions Standardizes the agent<>tool interface Conversational interface for every technical tool Model Context Protocol - MCP
  • 18.
  • 19.
    Out-of-the-box deployment Exposing functional tools notan API wrapper OpenMetadata MCP Server
  • 20.
    Out-of-the-box deployment Exposing functional tools notan API wrapper Native authentication & authorization OpenMetadata MCP Server
  • 21.
    Automate changes up anddown a data stack Combining with other MCP servers OpenMetadata MCP Server Use Cases Murat Migdisoglu, Data Architect & contributor Aimen Denche, Data Engineer @
  • 22.
  • 23.
  • 24.
    Learning Outcomes OpenMetadata isan open-source metadata platform provides discoverability, observability, and governance for every data asset in your organization to every one and every agent Model Context Protocol standardizes the interface between agents and tools, giving AI the right context to make the right decisions OpenMetadata MCP Server brings the context and governance of your full data stack to AI agents, allowing them to take actions the improve your data systems at scale
  • 25.
  • 26.
    Star us onGitHub https://github.com/open-metadata/OpenMetadata Join our Slack https://slack.open-metadata.org/ Follow us on X @open_metadata Collate is OpenMetadata aaS https://www.getcollate.io/ Thank you