All services Every engagement has a fixed, published price. Book a strategy call
Plain answers: what data governance is, what it is not, why AI and UK GDPR made it unavoidable in 2026, and what fixing it costs, with published fee floors.
Data governance has a definition problem. Ask five consultancies what it is and you get five frameworks, forty capabilities and a diagram that looks like a power station. Meanwhile the actual questions stay unanswered: who owns the customer record, why do two systems disagree about revenue, and who approved feeding that spreadsheet into an AI model. This guide gives the plain answers, what changed for 2026, and what fixing it costs.
The short answer. Data governance is the discipline of deciding who owns each important set of data, what good looks like for it, who may use it for what, how long it is kept, and being able to prove all of that when a regulator, auditor or AI programme asks. It is a set of decisions and habits rather than a tool you buy. Most organisations start with an independent assessment: with Intelance, a Data Governance QuickScan runs from £14,500 + VAT and shows where ownership, quality and evidence stand.
Strip the frameworks away and data governance answers five questions about every dataset that matters. Who owns it, meaning one named person is accountable when it is wrong. What good looks like, meaning the quality standard it must meet to be trusted. Who may use it, meaning access follows a rule rather than a request history. How long it lives, meaning retention and deletion happen on purpose. And how you prove it, meaning the answers exist as evidence rather than folklore. An organisation that can answer those five questions for its critical data is governed. An organisation with a thirty page policy and no answers is not.
It is not a catalogue purchase. Tools help once decisions exist, and an ungoverned catalogue is an expensive inventory of chaos. It is not the same as data management, which is the day to day engineering work of storing, moving and transforming data. Governance decides the rules and data management executes them. It is not a GDPR project either. UK GDPR compliance is one outcome of governed data, but privacy law covers personal data only, while governance covers everything the business runs on, including the financial and operational data that never touches a privacy register. And it is not a one off programme. Data changes daily, so governance is a running discipline with an owner, a rhythm and a budget.
Three pressures moved data governance from worthy to unavoidable. First, AI. Every AI initiative inherits the data underneath it, and the EU AI Act makes data governance an explicit legal duty for high risk systems under Article 10, covering training data relevance, representativeness and bias. Firms rolling out AI on ungoverned data are now taking legal risk as well as quality risk, which is why our AI governance practice treats data governance as a precondition. Second, accountability. UK GDPR expects organisations to demonstrate compliance rather than assert it, and boards increasingly ask for the same standard of evidence about data that they expect about money. Third, cost. Storage is cheap but bad data is expensive: duplicated records, contradictory reports and manual reconciliation are a permanent tax paid monthly, invisibly, in salaries.
| Question | The governed answer |
|---|---|
| Who owns the customer record? | A named business owner, with the authority and budget to fix it, rather than the IT department by default |
| Why do two systems disagree about revenue? | They cannot for long, because one source is designated as authoritative and the divergence is measured and worked |
| Who can see and use this dataset? | Access follows documented rules by role and purpose, reviewed on a schedule |
| Can we feed this data into an AI model? | A documented decision exists covering quality, rights, privacy and bias before the model sees it |
| Can we prove any of the above? | Yes, from a register and decision log kept as evidence, in the organisation’s own tools |
Start by measuring rather than by writing policy. An independent assessment shows where ownership, quality and evidence stand: our Data Governance QuickScan does this from £14,500 + VAT and ends with a scored register and a sequenced plan. Where the discipline needs building, Data Governance Foundations from £45,000 + VAT designs the operating model: owners, forums, standards and decision rights that survive delivery pressure. To run the discipline without hiring, a Virtual CDO from £7,500 + VAT per month provides the senior leadership, and data governance as a service from £9,500 + VAT per month runs the full function on one retainer. Where the need is one programme, such as an ERP replacement, migration or AI rollout, Project DG from £22,000 + VAT embeds the discipline in the programme and leaves when it ends. All fees are floors, agreed in writing before work starts, and everything produced lives in your own tools. The full practice is on our data governance consulting page.
AI governance decides which AI systems the organisation may run and under what controls. Data governance decides whether the data feeding those systems can be trusted and used lawfully. The two disciplines share a register mindset and an evidence habit, and in 2026 they increasingly share a regulator. The practical sequencing is that data governance questions surface within weeks of any serious AI governance work, which is why organisations planning AI adoption at any scale should assess their data first and their models second.
Deciding who owns each important set of data, what quality standard it must meet, who may use it for what, how long it is kept, and being able to prove those answers with evidence. It is a running discipline rather than a product or a one off project.
Governance decides the rules: ownership, standards, access and retention. Data management is the engineering that executes them: pipelines, storage, integration and tooling. Organisations commonly buy data management tools hoping to receive governance, and end up with well engineered chaos. Our full comparison of data governance and data management gives a test you can apply in a meeting and maps the common symptoms to the discipline that fixes each one.
No. UK GDPR compliance is one outcome of governed data, and it covers personal data only. Data governance covers everything the organisation runs on, including financial, operational and product data that privacy law never touches. Governed data makes demonstrating GDPR accountability far easier.
A named senior owner, with business data owners underneath for each critical dataset. In larger organisations that is a chief data officer. Where the estate does not justify a permanent hire, a fractional arrangement such as a Virtual CDO from £7,500 + VAT per month provides the same accountability.
With Intelance, an independent assessment runs from £14,500 + VAT, an operating model build from £45,000 + VAT, and a running function from £9,500 + VAT per month. Fees are floors and are agreed in writing before work starts.
An assessment reports within weeks. An operating model takes months to design and embed. The discipline itself is permanent, which is the honest answer most frameworks avoid: data governance ends when the data does.
Emmanuel Olatunji
Chief Enterprise Architecture Partner
Emmanuel founded Intelance in 2017 and has led architecture, AI governance, cyber and deal engagements across the UK, Europe, the Middle East and Africa. Every Intelance engagement runs under his direction, and every guide he publishes comes from that work.
See his record on LinkedIn About IntelanceThirty minutes with a senior adviser. A clear view of where to start, what it costs and what your team can run themselves.
Talk to a senior adviser
Data Governance
Data Governance
Data Governance