◆ Data & Analytics

What a data steward
really does.

6 tasks, each one witnessed by the sources that watched the job — and behind every one, a prompt you can use tonight.

6evidenced tasks
3systems it runs on
See the 6 tasks ↓Open in the atlasThe full career map →
This is what one task looks like here
Define and document data standards and business rules
Write the organisation’s canonical data standards and business rules f…1 sources agree

The shape of the day

tap a movement to see its tasks

Which one is you, right now?

Pick the moment · no score, no sign-up
Which moment is me right now?
Whichever you pick, the task behind it opens below.

The work, task by task

6 tasks
Hands on the work4
Manage metadata and data definitions+
Curate and approve metadata entries for the customer, product, and transaction domains: standardise column descriptions, map synonyms to preferred business terms, record data lineage for critical pipelines, and publish the updated definitions to the shared catalog for analyst use.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Ensure data compliance with regulations like GDPR+
Assess the customer and marketing data sets for GDPR exposure: inventory personal data elements, document lawful bases for processing, add retention and access restrictions to the data register, and send the compliance report with remediation steps to the DPO by Friday.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Implement and enforce data governance policies+
Roll out the updated data governance policy across the organisation: publish the policy with role-based responsibilities, enforce mandatory data stewardship sign-up, configure approval gates for new datasets, and report adoption metrics to the governance board in two weeks.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Collaborate with IT and business teams on data management+
Coordinate a cross-functional workshop with IT, product, and analytics to agree data ownership and SLAs for the sales pipeline, capture action items into the shared data register, assign remediation owners, and circulate the minutes and next steps within 48 hours.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Keep the record1
Define and document data standards and business rules+
Write the organisation’s canonical data standards and business rules for customer records, including field-level definitions, allowed formats, mandatory values, transformation logic for ingestion, and a versioned approval trail; publish the document to the data dictionary and notify the product and compliance leads for sign-off by next Wednesday.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Watch and assess1
Monitor data quality and resolve issues+
Run the daily data quality checks for sales and customer tables, triage any failing rules to identify root cause, create tickets for data owners with sample rows and timestamps, and confirm fixes before closing the tickets and updating the quality dashboard.
web
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?

What the work runs on

named inside the evidenced tasks
4 tasksData catalogsstores and publishes canonical definitions and versioned documentation for business users
3 tasksMetadata management toolscaptures lineage and transformation logic tied to each element
2 tasksData quality monitoring platformsexecutes scheduled rules, alerts on failures, and tracks remediation status
Browse every evidenced software task →

The same task, four heights

this page is height one
ExecuteDo today's task, with fewer mistakesyou are here → ImproveMake it easy for the next person to acceptin the atlas → DecideWork out the right move when it is unclearin the atlas → BecomeLearn the pattern so it stops coming backin the atlas →

Can AI actually do this job?

the honest answer

It can

where it genuinely helps
  • Explain the theory behind the work
  • Draft, tidy and structure your writing
  • Rehearse a hard conversation before you have it
  • Build a study plan that fits your gaps

It cannot

where it stops, completely
  • Be in the room where a data steward actually works
  • Carry the responsibility when the call is wrong — that weight stays yours
  • Notice what no one wrote down: the hesitation, the thing left unsaid
  • Live with the outcome

Where the evidence lives

open any of it yourself

Close to this work

12 nearby
Data & AnalyticsFinancial Analyst26 evidenced tasks Data & AnalyticsExperimentation Analyst26 evidenced tasks Data & AnalyticsData Governance Analyst26 evidenced tasks Data & AnalyticsR Analyst26 evidenced tasks Data & AnalyticsCrm Analyst26 evidenced tasks Data & AnalyticsWeb Analyst26 evidenced tasks Data & AnalyticsGrowth Analyst26 evidenced tasks Data & AnalyticsProduct Analyst25 evidenced tasks

Questions people actually ask

You spend most of your day checking that data follows the rules your company set. That means using data catalogs and metadata management tools to look up data definitions, then reviewing data quality dashboards to find missing or wrong values.

You also write or update business rules and data standards documents, meet with IT and business teams to solve data issues, and log fixes in whatever ticket system your company uses. Expect regular reviews and ad-hoc firefighting when a data feed breaks.

Common tools are data catalogs (like Collibra or Alation), metadata management tools, and data quality monitoring platforms (like Great Expectations or Informatica Data Quality). You’ll use the catalog to find what a field means, the metadata tool to track lineage (where data came from), and the quality platform to run checks and alerts.

You also often use spreadsheets, SQL editors, and a ticketing system (Jira or ServiceNow) to record issues and coordinate fixes with IT and business owners.

First, monitor quality dashboards for alerts: high null rates, duplicates, or out-of-range values. The data quality platform will show which dataset and which rule failed. Next, check the metadata and lineage in the data catalog to find the source system and owner.

Then open a ticket, describe the failed rule and evidence, and work with the source owner (often IT or the business unit) to correct the source or apply a transformation. Log the resolution and update the business rule if needed.

AI can help spot patterns, suggest data mappings, or auto-classify fields in a data catalog, but you must not trust it blindly. Use AI suggestions as a starting point, then validate with metadata, samples, and a human review before updating definitions or rules.

Never use AI to make compliance decisions alone—always check outputs against your GDPR policies and data governance rules. Keep records of AI suggestions and who approved changes.

A data steward focuses on rules, definitions, and quality — the policy and ownership side. You maintain metadata, set data standards, and coordinate fixes. Data engineers build and run pipelines that move and transform data; they change code and infrastructure.

Data analysts consume the cleaned data to produce reports and insights. Stewards sit between engineers and analysts: you enforce the standards engineers implement and ensure analysts understand the data definitions in the catalog.

Start with SQL to query data and spot quality issues. Learn one data catalog (Collibra or Alation) and one data quality tool (Great Expectations or a similar platform) so you can read metadata and run checks.

Also practice writing clear business rules and documentation, and get comfortable coordinating with IT and business teams. Basic knowledge of GDPR and other data compliance rules is essential.

Common numbers are data quality KPIs: percent complete (target 95%+), duplicate rate (goal under 1% for key identifiers), and timeliness (e.g., 99% of daily feeds arrive by 07:00). You’ll track rule failure counts and mean-time-to-resolution for incidents.

For metadata, you might measure catalog coverage (percent of critical datasets with definitions; target 100% for high-value data) and data ownership coverage (percent of datasets with assigned owners). These numbers guide your priorities.