Data Asset Management (DAM) capability playbooks
Have your sayHelpSite map

Learning journeys Level: Foundation

Learn the playbooks

Short modules that build from the basics of the policy to running all three capabilities as one yearly cycle. Each module lists what you will be able to do, the activities, and a quick check of understanding.

  • 19 modules
  • 470 minutes
  • 3 levels

In brief

  • 19 modules, about 470 minutes in total, at three levels.
  • Each module has outcomes, a short set of activities and a check of understanding.
  • 1 themed pathway, on AI-ready data stewardship, joins the modules to sister projects.
  • Follow the journey for your role, or pick any module.

Levels

Journeys by role

Pathways

Themed routes through the modules. Each one says what it covers in other projects, what is still missing and the strongest case against it.

6 modules · about 160 minutes · links checked 2026-11-14

AI-ready data stewardship

Six modules for owners, stewards and catalogue managers who are asked whether a critical data asset can be used for AI. They build from what AI-ready means, through metadata and linked data, to governing AI use, and join this portal to five sister projects so you do not have to find them yourself.

Activities done0 of 31 done

The route

  1. M12 What makes a critical data asset AI-ready Foundation 20 minNot started
  2. M13 From FAIR to AI-ready metadata Practitioner 30 minNot started
  3. M14 Start from the question the data must answer Practitioner 20 minNot started
  4. M15 Connecting data to AI with linked data, knowledge graphs and MCP Advanced 35 minNot started
  5. M16 Making public data AI-ready together Practitioner 25 minNot started
  6. M17 Governing AI use of critical data in practice Advanced 30 minNot started

How it lines up with Data Stewardship at a Time of AI, a deep-dive series

Data Stewards Academy (The Data Tank and The GovLab). Five weekly live sessions of 90 minutes, Wednesdays 10 March to 7 April 2027, as published on 14 November 2026. An independent course, not connected with this portal. The rows below show where this pathway prepares you for, or follows up, each session. Check dates and details on the course page. Course page.

Course theme, pathway modules and how well they cover it
Course themeModulesCoverNote
What does 'AI-ready' actually mean?M12StrongGrounded in the GDS guidelines and the AI-ready supplement.
From FAIR to AI-readyM13 M14PartialFAIR and metadata are covered well. Demand-led problem definition is only started in M14.
Connecting data to AI: MCP and knowledge graphsM15StrongUses working tools: this portal's graph and MCP server, NDL PubSec MCP and DCAT-AP-UK.
Making public data AI-readyM16PartialDiscoverability and assurance are covered. Participatory and community-led stewardship is a gap.
AI-ready data in actionM17StrongRunnable rules, transparency duties and the AI governance toolkit.

What is still missing

  • PG-01 Participatory and community-led stewardship Open
    No project in this estate yet gives guidance on data cooperatives, community charters or involving the public in decisions about AI use of their data. M16 points to the inclusion questions only.
  • PG-02 Questions-led demand for data Started
    M14 asks you to write down the question an AI use serves. There is no method yet for gathering and prioritising questions from users and the public.
  • PG-03 International comparison Open
    The pathway is UK-focused. It does not yet compare how other governments prepare public data for AI.

The case against

Another learning route spreads people thin. Stewards could go straight to the GDS guidelines and an external course, and a pathway that links to five prototypes inherits their churn and their unendorsed status.

Our answer

The pathway adds no new guidance of its own. It reuses the GDS guidelines, the existing AI-ready supplement and sister projects, labels each one's status, and keeps the checks of understanding to facts you can trace. If a sister project moves or closes, the weekly health run reports the broken link.

Where the pathway sends you, and each source's status

All modules

Learn it, keep it, share it

Learning sticks when you use it, come back to it and explain it to someone else. These tips help you learn from the playbooks and the sources they cite, keep what you learn, share it, and count it towards your role, your career and continuing professional development (CPD).

Learn it

  • Start from a real asset you look after. Read each module with that asset in mind, not in the abstract.
  • Do the activities, not only the reading. Each one produces something you need for the return anyway.
  • Read the source behind a point when it matters: every fact links to the GOV.UK page or standard it comes from.
  • Use the glossary and tooltips the first time a term is new. Hover or focus on a dotted term to see what it means.
  • Follow the journey for your role first, then pick other modules as your work needs them.

Keep it

  • Try the check of understanding before you reread the module. Recalling an answer helps you keep it more than reading again.
  • Come back to a module a week later and a month later. The page reminds you when a module is due for a refresher.
  • Write one sentence after each module: what you will do differently on Monday.
  • Keep your filled templates. They are a record of what you did and a reminder next year.
  • Use your progress record as a checklist before each annual return.

Share it

  • Explain one module to a colleague in ten minutes. Teaching it is the quickest way to find what you have not understood.
  • Run a lunch-and-learn on a playbook with your owners and stewards, using the module activities as the agenda.
  • Post a question or a lesson in the cross-government data management community, and in the playbooks' discussions.
  • Suggest a fix when something here is unclear: every page has a link to report a problem or suggest a change.
  • Share your filled templates inside your organisation so the next person starts from yours.

Tips for your role

Learning tips by role
RoleTip
Data ownerTake M01 and M02 first. Your job is decisions, not data entry: know your assets, your steward and your review date.
Data stewardTake M04 to M08. Keep a short list of the questions users ask you; it tells you what to fix first.
Catalogue managerTake M05 and M06 and keep the metadata guidance open while you work. Check one record a week against it.
Data quality lead or analystTake M07 and M08. Score one real issue with the issues framework before reading further.
Chief data officerTake M01, M09 and M11, then the leader route. Ask for the maturity baseline in your first month.
Lead policy coordinatorTake M01 to M03 and M09. Put the yearly cycle in the calendar before anything else.

Refreshers due

Modules you finished a week or a month ago come back here for a quick look at the check of understanding.

Finish a module to see refreshers here.

Your role, career and certifications

What you learn here counts towards your role, your profession and your next job. The modules map to roles in the Digital and Data Capability Framework and to government professions, so you can use them as evidence in a development conversation, a job application or a promotion.

  1. Find your role, or the one you want, on the Professions and skills page, and note its skills.
  2. Pick the modules that build those skills, and add them to your development plan.
  3. Record each module you finish, what you did with it and the evidence, in your CPD log.
  4. Use the record in your performance review and in Success Profiles applications.
  5. When you are ready, choose a certification that fits your role and aim for it.

Government Digital and Data Profession Capability FrameworkV: Verified against the primary source on 2026-11-05 · Government professionsV: Verified against the primary source on 2026-11-05 · Success ProfilesV: Verified against the primary source on 2026-11-05 · Data skills for allV: Verified against the primary source on 2026-11-05 · Government Campus prospectusV: Verified against the primary source on 2026-11-05 · Government Knowledge and Information Management ProfessionV: Verified against the primary source on 2026-11-05 · Analysis Function career frameworkV: Verified against the primary source on 2026-11-05 · SFIA 9: the global skills and competency framework for a digital worldV: Verified against the primary source on 2026-11-05

Certifications to consider

Independent schemes, listed in no order and not endorsed. Each sets its own rules; check them on the scheme's own pages.

Certifications
CertificationWho it fitsCPD and renewalSource
Certified Data Management Professional (CDMP)
DAMA International
Data management professionals across all knowledge areas, including governance, metadata and quality.Holders recertify every three years; schemes commonly cite 120 hours of continuing education per three-year cycle. Check the current rules on the scheme's own pages.S: Found by web search; the site blocked automated checking, so open it before relying on itCDMP: Certified Data Management ProfessionalS: Found by web search; the site blocked automated checking, so open it before relying on it
BCS Practitioner Certificate in Data Management Essentials
BCS
Business analysts, managers and project managers who need a grounding in data management in an organisation.A certificate gained by exam. Log the study hours as CPD.S: Found by web search; the site blocked automated checking, so open it before relying on itBCS Practitioner Certificate in Data Management EssentialsS: Found by web search; the site blocked automated checking, so open it before relying on it
CILIP professional registration (certification, chartership and knowledge management chartership)
CILIP
Knowledge and information management professionals, including records and catalogue roles.Revalidation asks for at least 20 hours of CPD a year and a short statement. Check the current rules on CILIP's pages.S: Found by web search; the site blocked automated checking, so open it before relying on itCILIP professional registrationS: Found by web search; the site blocked automated checking, so open it before relying on it
Royal Statistical Society professional development
Royal Statistical Society
Statisticians and analysts working on data quality and official statistics.Professional status includes keeping a CPD record. Check the society's pages.S: Found by web search; the site blocked automated checking, so open it before relying on itRoyal Statistical Society professional developmentV: Verified against the primary source on 2026-11-05

Counting it as CPD

Your learning record

A file of the modules you have finished in this browser, with their time, level and outcomes, and the date you finished. Add your own notes and use it as the start of your CPD log.

Download my learning record (CSV)

CPD and learning log

CPD and learning log (T-LRN-1)

Business as usual

  • Date
  • Activity (module or source)
  • Kind (course or reading or practice or teaching)
  • Hours
  • What I learned
  • How I applied it at work
  • Evidence (location)
  • Skill or role it supports
  • Certification or scheme it counts towards
  • Refresher due

Download this template (CSV) · All templates (CSV)

Modules in full

M01 · Foundation · 15 minutes

The DAM Policy - what it asks of you

Governance Cataloguing Quality
Activities done0 of 4 done

By the end you will be able to

  • Explain who the policy applies to and what is in scope
  • Name the core duties: identify and catalogue assets, report them, name an owner and manage quality
  • Find the requirement behind any step in these playbooks

Activities

  1. Reflect. Which of the three capabilities is least developed in your organisation today?

Check your understanding

Which organisations must follow the policy?

Show the answer

UK government departments and their arm's length bodies. It applies to UK government departments and their ALBs. It is optional for public corporations and does not apply to councils.

What must every identified data asset have?

Show the answer

A named data owner and a DQAP or equivalent. The policy requires a named data owner and a data quality action plan, or an equivalent mechanism, for every identified asset.

M02 · Foundation · 15 minutes

Roles and accountability

Governance
Activities done0 of 4 done

By the end you will be able to

Activities

  1. Do. List your critical data assets and the role that should own each one.
  2. Reflect. Does every owner in your list have a named successor?

Check your understanding

Which role is accountable for an asset's meaning, content and quality?

Show the answer

Data owner. The data owner is the senior leader accountable. The steward runs the asset day to day; the custodian runs the technical system.

What must happen when a data owner leaves?

Show the answer

A successor is ready and the metadata is updated. The guidance asks you to tell the coordinating team, have a successor ready and update the metadata.

M03 · Practitioner · 30 minutes

Setting up governance with five recipes

Governance
Activities done0 of 4 done

By the end you will be able to

  • Run the Mandate, Frame, Assign, Operate and Assure recipes in order
  • Set up a decision log and an exception register
  • Collect evidence continuously for the GDS return

Activities

  1. Reflect. Where will your evidence folder live, and who keeps it current?

Check your understanding

What makes an exception different from a policy change?

Show the answer

It has an owner and an expiry date. An exception is time-bound with an owner and an expiry date. Without an expiry it is an undocumented policy change.

When should you assemble evidence for the return?

Show the answer

Continuously during Operate. Collecting evidence each time a decision is logged turns the return into collation rather than reconstruction.

M04 · Practitioner · 25 minutes

Finding and designating critical data assets

Cataloguing
Activities done0 of 4 done

By the end you will be able to

  • Build a candidate list from existing registers
  • Apply the scope tests and the seven purposes
  • Draw an asset's boundary and record why

Activities

  1. Reflect. Have you looked at data you acquire from others, not only data you provide?

Check your understanding

If you designate a distribution of a data set, what else should you designate?

Show the answer

Its source data set. The guidance asks you to designate the source data set as well.

Which of these is one of the seven purposes?

Show the answer

Official statistics. Official statistics is one of the seven. The statistics themselves are not designated, only the data used to produce them.

M05 · Practitioner · 30 minutes

Describing assets with metadata others can use

Cataloguing
Activities done0 of 4 done

By the end you will be able to

  • Record the metadata properties in the GOV.UK guidance
  • Write titles and descriptions in plain English
  • Map properties to the Metadata Exchange Model

Activities

  1. Do. Ask someone outside your team to read the record and say what the data is for.
  2. Reflect. Which properties did you have to mark 'unknown'? Who could supply them?

Check your understanding

How should you record a date in a catalogue record?

Show the answer

In ISO 8601 format. The guidance asks for ISO 8601, for example 2020-07-14.

What should you put when a property has no value?

Show the answer

'unknown' or 'not applicable'. Use 'unknown' or 'not applicable' rather than leaving a blank.

M06 · Practitioner · 20 minutes

Cataloguing, harvesting and reporting to GDS

Cataloguing
Activities done0 of 3 done

By the end you will be able to

  • Expose records so others can harvest them
  • Produce the return from catalogue records
  • Plan the annual review

Activities

  1. Do. Check that every link in one record works from outside your network.

Check your understanding

How often must metadata be reviewed?

Show the answer

At least every 12 months. The policy asks for metadata to be maintained and reviewed annually.

Who assures the returns of arm's length bodies?

Show the answer

Their department. Departments coordinate and assure the returns of their ALBs.

M07 · Foundation · 15 minutes

Data quality foundations

Quality
Activities done0 of 3 done

By the end you will be able to

  • Name the five principles and six dimensions of the Government Data Quality Framework
  • Explain what a DQAP is for
  • Recognise when quality differs between groups of people

Activities

  1. Reflect. Which dimension matters most for your most important decision?

Check your understanding

Which is one of the six dimensions?

Show the answer

Timeliness. The six are completeness, uniqueness, consistency, timeliness, validity and accuracy.

Why break quality results down by group?

Show the answer

So poor quality for some people is not averaged away. An average can hide poor quality for some groups, which can exclude them from services.

M08 · Practitioner · 35 minutes

Building a DQAP and scoring issues

Quality
Activities done0 of 4 done

By the end you will be able to

  • Write rules with targets from the issues framework
  • Measure a baseline and log issues
  • Prioritise issues and set resolution dates

Activities

  1. Do. Score an issue with the issues framework calculator. (Public Sector Data Quality Commons)
  2. Reflect. Which issue would you fix at source rather than clean downstream?

Check your understanding

How does the issues framework set a rule's target range?

Show the answer

From importance of purpose times impact. Importance of purpose multiplied by impact gives a combined score, which sets the target range.

In the framework's example timescales, when is a critical issue resolved?

Show the answer

Immediately. The framework's example gives Critical as immediate; High, Medium and Low within 1, 3 and 6 measurements.

M09 · Advanced · 20 minutes

Running the three capabilities as one yearly cycle

Governance Cataloguing Quality
Activities done0 of 3 done

By the end you will be able to

  • Use the hand-offs so each fact is recorded once
  • Plan one annual review per asset
  • Feed assurance lessons into the next mandate

Activities

  1. Do. Pick one asset and check the owner is the same in the register, the record and the DQAP.

Check your understanding

Where should the owner and successor for an asset be recorded first?

Show the answer

In the ownership register, then reused. The ownership register is the source; the catalogue record and the DQAP reuse it (IL-01, IL-02).

M10 · Foundation · 20 minutes

Inclusive data practice

Governance Cataloguing Quality
Activities done0 of 3 done

By the end you will be able to

  • Apply equality and inclusion checks in each capability
  • Make data quality statements and records accessible
  • Involve people the data describes

Activities

  1. Reflect. Whose voice is missing when you decide what 'good enough' means?

Check your understanding

Which duty asks public bodies to consider equality in their work?

Show the answer

The public sector equality duty. The public sector equality duty in the Equality Act 2010 applies to public authorities in their functions.

M11 · Practitioner · 30 minutes

Embedding the policy in delivery and change

Governance Cataloguing Quality
Activities done0 of 4 done

By the end you will be able to

  • Name the capabilities the policy relies on but does not state
  • Build DAM checkpoints into a business case and an assurance plan
  • Plan a legacy exit and a handover to business as usual

Activities

  1. Reflect. Which critical asset would be left without an owner if its project closed tomorrow?

Check your understanding

Which guidance makes a register of data assets its second principle?

Show the answer

Managing legacy technology. The GOV.UK guidance on managing legacy technology asks organisations to build a complete and accurate register of their data assets.

How does the capability model flag a need the DAM Policy text does not state at all?

Show the answer

Inferred. Inferred means the need is not in the text and comes from good practice and adjacent rulebooks.

M12 · Foundation · 20 minutes

What makes a critical data asset AI-ready

Governance Cataloguing Quality AI-ready data stewardship
Activities done0 of 4 done

By the end you will be able to

  • Name the four pillars in the GDS guidelines on AI-ready datasets
  • Decide whether an asset needs the AI-ready supplement
  • Explain why AI-readiness is a stewardship question as well as a technical one

Activities

  1. Reflect. If someone asked tomorrow to train a model on your most important asset, who would say yes or no, and on what evidence?

Check your understanding

Which of these is one of the four pillars of AI-ready data in the GDS guidelines?

Show the answer

Data and metadata quality. The four pillars are technical optimisation; data and metadata quality; organisation and infrastructure context; and legal, security and ethical compliance.

When does the AI-ready supplement ask you to publish an algorithmic transparency record?

Show the answer

Where a tool supports decisions about people. The supplement asks for a record where an algorithmic tool supports decisions about people, using the Algorithmic Transparency Recording Standard.

M13 · Practitioner · 30 minutes

From FAIR to AI-ready metadata

Cataloguing AI-ready data stewardship
Activities done0 of 5 done

By the end you will be able to

  • Explain the four FAIR principles in plain English
  • Add the AI-ready catalogue fields to a record
  • See how the proposed DCAT-AP-UK profile extends the xGov MEM for AI use

Activities

Check your understanding

What do the letters FAIR stand for?

Show the answer

Findable, accessible, interoperable, reusable. The FAIR Guiding Principles (Wilkinson and others, 2016) are findability, accessibility, interoperability and reusability, for people and for machines.

Why does the AI-ready supplement ask for a version on the record?

Show the answer

So a training set can be reconstructed exactly. The supplement asks for dcat:version so a training set can be reconstructed exactly.

Which metadata model is endorsed for UK government today?

Show the answer

The xGov MEM. The UK Cross-Government Metadata Exchange Model (xGov MEM) is endorsed by the Data Standards Authority Steering Board. DCAT-AP-UK is a proposal for review.

M14 · Practitioner · 20 minutes

Start from the question the data must answer

Governance Cataloguing AI-ready data stewardship
Activities done0 of 5 done

By the end you will be able to

  • Write down the decision or question an AI use of an asset serves
  • Test that question against the seven purposes in the policy
  • Record out-of-scope uses before anyone asks for them

Activities

  1. Do. For one asset, write one sentence each: the decision it informs, who makes it, and what goes wrong if the data is wrong.
  2. Reflect. Who outside your team, including the people the data describes, should help frame the question?

Check your understanding

What does the AI-ready supplement ask you to state before AI use begins?

Show the answer

The task, whether outputs affect individuals and which uses are out of scope. Its first governance check is to record the AI task, whether outputs affect individuals and which uses are out of scope, such as automated decisions.

Which of these is one of the seven purposes?

Show the answer

Official statistics. Official statistics is one of the seven purposes in the policy.

M15 · Advanced · 35 minutes

Connecting data to AI with linked data, knowledge graphs and MCP

Cataloguing AI-ready data stewardship
Activities done0 of 6 done

By the end you will be able to

  • Explain the RDF triple and why linked data helps machines read metadata
  • Find and reuse this portal's knowledge graph and vocabulary files
  • Describe how an MCP server lets an AI assistant check metadata against the xGov MEM

Activities

  1. Read. Browse the knowledge graph and thesaurus of the proposed DCAT-AP-UK profile. (DCAT Application Profile for the UK (DCAT-AP-UK))
  2. Reflect. Which of your catalogue's fields would an assistant need to answer 'can I use this for training?' without guessing?

Check your understanding

What are the three parts of an RDF triple?

Show the answer

Subject, predicate and object. The RDF 1.1 Primer describes every statement as a triple of subject, predicate and object.

What does the NDL PubSec MCP tool metadata_validate do?

Show the answer

Checks a metadata record against the xGov MEM and lists errors and warnings. Its README says it checks any metadata record against xGov MEM and lists errors and warnings with their paths. All its tools are read-only.

M16 · Practitioner · 25 minutes

Making public data AI-ready together

Governance Quality AI-ready data stewardship
Activities done0 of 5 done

By the end you will be able to

  • Look for existing work before building, across public sector code and catalogues
  • Involve the people the data describes when you judge whether it is fit for AI
  • Use the shared assurance and community routes rather than working alone

Activities

  1. Reflect. Whose consent, trust or voice would change your answer to 'is this asset AI-ready'?

Check your understanding

What does the GitHub Repos Asset Catalogue ask you to do first?

Show the answer

Look for work that already exists before you build. Its motto is 'Before you build it, look for it': find public sector code that already exists and decide whether you can reuse it.

Where should harvested catalogue records from government end up?

Show the answer

The National Data Library gateway at data.gov.uk. data.gov.uk is the cross-government gateway to public data and the destination for harvested catalogue records.

M17 · Advanced · 30 minutes

Governing AI use of critical data in practice

Governance Quality AI-ready data stewardship
Activities done0 of 6 done

By the end you will be able to

  • Run the AI-ready quality rules on a sample file
  • Apply the transparency and ethics duties that go with AI use
  • Plan who monitors drift and bias once AI use begins

Activities

  1. Reflect. If a model trained on your asset started giving worse answers for one group, how would you find out, and who would act?

Check your understanding

Why does rule AI-03 check that each record appears once?

Show the answer

So training and test sets do not overlap. AI-03 is a uniqueness rule: each record appears once, so training and test sets do not overlap.

Which operating model check does the AI-ready supplement include?

Show the answer

Someone owns quality, drift and bias monitoring. Its operating model checks include that someone owns quality, drift and bias monitoring, and that training sets can be reconstructed exactly.

M18 · Foundation · 25 minutes

Leading data asset management - your first year

Governance Cataloguing Quality
Activities done0 of 4 done

By the end you will be able to

  • Set Must, Should and Could aims for your organisation and the outcomes each brings
  • Plan the first 30, 60 and 90 days and the 6 and 12 month checkpoints
  • Choose the maturity model that fits each aim and measure progress against a baseline

Activities

  1. Reflect. Which strand would you most need to move by the six-month checkpoint, and what would stop it?

Check your understanding

Which aim is a Must for every organisation in scope?

Show the answer

Every critical asset is identified, owned, catalogued, has a quality plan and is reported. The DAM Policy asks this of critical data assets. Covering every asset is a later aim, and AI readiness is a Could.

Around which maturity level does the DAM Policy minimum for critical assets sit?

Show the answer

Learning. The crosswalk places the policy minimum around Learning: processes defined and repeatable for critical assets.

M19 · Practitioner · 30 minutes

Buying tools for data asset management without lock-in

Governance Cataloguing Quality
Activities done0 of 4 done

By the end you will be able to

  • Pick the requirements that fit the kinds of tool you need, with priorities
  • Run a request for information and a proof of concept with your own data
  • Ask for contract clauses that keep your data movable

Activities

  1. Reflect. If your current tool's contract ended next month, could you take everything with you?

Check your understanding

How should a Must requirement be treated when scoring responses?

Show the answer

Pass or fail. A response that does not meet a Must fails, however it scores elsewhere.

What is the best evidence that a tool will let you leave?

Show the answer

A full export and re-import in the proof of concept. A round trip with your own data shows what is lost; a list or a statement does not.

Page owner: the portal's service owner (a role) · Published 5 November 2026 · Last updated 2026-11-06 · Next review 2027-02-05

Was this page useful? Tell us Tell us privately, without GitHub