---
title: "MDLH: High value use cases"
url: "https://atlan.com/demos/mdlh-high-value-use-cases/"
excerpt: "Explore the six high-impact use case areas where teams consistently find the most value from day one."
description: "Cut through the “where do I even start?” question with a guided tour of the use cases that deliver real results: metrics consistency, context distribution for AI and source systems, proactive impact analysis, root cause investigation, cost optimization, tag compliance, and platform adoption analytics. Each one maps to a specific business problem your team is already dealing with — now you’ll know exactly which query to reach for."
format: "Video"
duration: "PT6M58S"
video: "https://videos.ctfassets.net/nwa1c00rtgxb/4QatpRdIdghyF0xDoU2Thz/25cf73e0cca9ebf54447bb79f5b281d9/MDLH_-_High_value_use_cases.mp4"
thumbnail: "https://images.ctfassets.net/nwa1c00rtgxb/4maCIeieWLQjwFllJjzzAO/abbf44af162eb19faba6547958950cde/mdlh-high-value-use-cases.webp"
content_purpose: ["Product Overview"]
target_persona: ["General", "Data Engineer", "Data Analyst", "Data Governance Lead"]
journey_stage: ["S2 - Discovery", "S3 - Solution Design", "S4 - Business Case", "C1 - Onboarding", "C2 - First Value", "C3 - Adoption"]
use_case_context: ["Training"]
product: ["Enterprise Data Graph - Integration"]
published: "2026-03-13"
updated: "2026-06-02"
content_type: "video transcript"
transcript_source: "sheet"
---

# MDLH: High value use cases

Transcript of the video at https://atlan.com/demos/mdlh-high-value-use-cases/

The Metadata Lakehouse is live in your workspace but Most teams have no idea where to start. Here are the use cases that consistently drive the most value — right away. Every time we show the Metadata Lakehouse to a new team, someone asks: that's powerful — but where do I even start? It's the right question. Not every query is equal. These are the key areas and use cases where teams consistently find the most value, Here’s how to think about them. The Metadata Lakehouse is an open SQL surface — powerful enough to support any metadata use case our team needs to tackle.

The key is knowing where to start. Here are the core use case areas where teams consistently find the most value. We'll look at what problem each one solves, what kind of query we'd build, and what the output could look like. The first is Metrics consistencyThis use case covers glossary analysis and export, ensuring our metrics remain consistent. If our organization defines revenue, active users, or churn in multiple places, or if different teams are working from slightly different versions of the same term, the metadata lakehouse gives us a way to query our entire glossary in SQL.

We can see all terms, their parent glossaries, categories, and which assets they're assigned to. That makes it straightforward to find duplicates, compare definitions side by side, and batch-export the output for stewardship review or bulk updates. That's the metrics and glossary use case, ensuring that you have a single source of truth that you can rely on. The next is context distribution — getting our enriched metadata out of Atlan and into the systems that depend on it. There are two destinations for that context. The first is AI. Before we can feed metadata to Snowflake Cortex, a RAG pipeline, a vector search system, or any other AI engine, we need to know which assets are actually well-documented.

The Lakehouse lets us run a completeness scorecard across our entire estate — description coverage, tag coverage, certified assets, owned assets — broken down by domain or asset type, so we know exactly where to focus enrichment sprints before we export. The second destination is source systems. Once the metadata is enriched in Atlan, the Lakehouse is how we distribute it back — pushing descriptions, certifications, and business context to dbt, data dictionaries, upstream tools, or any system that needs Atlan as its source of truth. Set it up on a cadence and the sync runs automatically.

One SQL surface, two directions. This is where the Lakehouse genuinely outpaces the API for both use cases — no pagination, no rate limits, a single SQL query across our entire estate whether we are building an AI context layer or keeping source systems in sync. That's the context distribution use case — AI agents and source systems, served from the same SQL surface. Next is investigating data issues and preventing them — through root cause and impact analysis. Building trust with data consumers isn't just about fixing problems when they arise; it's about being proactive enough to prevent them in the first place.

But when issues do occur, speed of resolution matters. And as more organizations extend their data consumers to include AI agents, bad data becomes an even faster trust-killer — incorrect outputs trace straight back to the pipeline that fed them. The Metadata Lakehouse addresses both sides of this. When we're planning a change, we can query the full downstream impact before we make it — surfacing every dependent asset, identifying key stakeholders, and giving us what we need to make decisions confidently. That small column rename doesn't have to become the reason the CRO's Monday morning dashboard breaks.

And when something is already broken, we can run the same kind of query in reverse — tracing upstream from a troubled asset to find where the issue originates. Instead of clicking through lineage hop by hop, we get a structured result with the indicators we need to pinpoint root cause quickly. In both directions, the Lakehouse returns exactly the metadata that matters for our situation — whether that's ownership, certification status, or a custom field specific to our organization — and gives us enough context to act on it. NExt is keeping our data estate lean and clean through cost optimization.

Any asset with no upstream lineage and no downstream lineage is an orphan — it's not feeding anything and nothing is feeding it. Combined with usage metadata — last query date, query frequency, storage size — MDLH gives us the data we need to make confident deprecation decisions. We're not deleting assets based on gut feel — we're querying for evidence that something has been unused for six months and has no dependencies. A single query surfaces the full lineage chain and last-used timestamps for every asset in scope. From cost to compliance — the fifth use case is making sure our confidential assets carry the right tags all the way through the lineage chain.

Tagging an asset as PII or Confidential in Atlan is one thing. Knowing whether that tag has propagated to all downstream consumers — through multiple lineage hops, across multiple steps in the lineage graph — is another. MDLH lets us calculate tag inheritance percentages at each hop and identify downstream assets where coverage has dropped. If we have 200 downstream tables derived from a tagged source and only 140 carry the tag, that gap is a compliance risk. MDLH surfaces it before our auditors do. A single query returns every downstream asset that should carry the tag — and flags the ones that do not.

And one more — this one is unique to Atlan itself and surprises most teams: platform analytics. We may be asked who is using Atlan, who are our power users, how is our adoption?the Metadata Lakehouse lets us query our own platform telemetry directly. Daily, weekly, and monthly active users. We can build a stickiness ratio that tells we whether Atlan is a daily habit or an episodic tool. Engagement tiers that classify every user as Power, Heavy, Light, or Dormant each month. And a power user ranking that surfaces our internal champions by consistency and breadth of use.

This is behavioral data, not just access data — the difference between knowing who has a license and knowing who's actually getting value. A targeted query returns daily, weekly, and monthly active users alongside the assets they are engaging with. Each of these use cases is a starting point — the Lakehouse gives us the SQL surface to build on from here. Every team runs into different metadata problems — the Lakehouse gives us the SQL surface to address them, whether that means automating governance actions or building an exportable context layer for AI agents.

If you want to start with the metadata completeness walkthrough — which allows you to understand your metadata enrichment, providing the context you need to power your AI agents and data initiatives — that walkthrough is linked right here.