How Sage Built an Award-Winning Data & AI Operating Model Across 2M Customers and 26 Countries

For most enterprises, the question is no longer whether to invest in AI — it is whether the underlying data is ready to be trusted at scale. Sage, a global leader in AI-powered business software serving 2M+ customers across 26 countries, built its answer around a deceptively simple premise: treat data as a product, govern it as a portfolio, and let AI inherit that discipline. DataGalaxy is the platform that made this possible.

Challenges

Like many large organisations, Sage faced a familiar set of frictions: duplicated analytics requests, inconsistent definitions of core business terms, and an ever-growing backlog of reports with no clear owner. These translated into six questions the team could no longer ignore:

• Trust: How do you get stakeholders from different product lines and regions to trust a centralised process?
• Consistency: How do you create a standardised delivery methodology for evaluating value and prioritising all data use cases?
• Business Outcome: How do you track and recognise the financial and non-financial outcomes of data use cases to ensure measurable value?
• Stakeholder Alignment: How do you bring key stakeholders together in a single place to drive data deliverables?
• Compliance: How do you create trusted, tested data products that comply with relevant governance and regulation?
• Risk Management: How do you assess and track risk across the full delivery lifecycle?

These questions became the catalyst for a new operating model — and the decision to build it on DataGalaxy.

Why DataGalaxy

Sage needed more than a data catalog. They needed a platform that could channel every data and AI request through a consistent evaluation process, connect every asset to the business outcome it served, and give leadership a live investment view of the entire portfolio. DataGalaxy provided all of that.

Three capabilities made DataGalaxy the right fit: a single environment to manage use cases end to end; governance standards enforced through metadata, not policy documents; and value lineage — the ability to trace every report, data product, and AI product back to the business initiative that justified it.

On that foundation, Sage built the Data Engagement Model (DEM) — a delivery framework recognised as a finalist at the British Data Awards in 2024 and 2025. It is the single channel through which every data and AI use case at Sage now flows. The shift the DEM delivers is threefold:

• From requests to use cases: Every business ask is screened, deduplicated, and prioritised against existing assets before any build begins.
• From projects to products: Delivery is structured around reusable, governed data products rather than one-off reports.
• From technology-first to value-first: Time-to-value is validated upfront, not assumed after delivery.

Data as a Product

Central to DataGalaxy’s role at Sage is making the data-as-a-product principle operational, not theoretical. Sage defines a data product as “a business-focused, reusable, and consumable artefact that delivers a defined business outcome through data.” Three non-negotiables anchor the discipline and prevent the model from collapsing back into bespoke project work:

• Reusable: Data products are validated and trusted by the business so they can serve multiple needs without rebuilding from source. Reuse is treated as a measurable outcome, not an aspiration.
• Value-Driven: A consistent methodology evaluates and prioritises every use case against business value, ensuring scarce delivery capacity flows to the highest-leverage initiatives first.
• Dynamic: Every data product follows an iterative lifecycle — ideation, build, adoption, continuous improvement, and eventual retirement — with quality and trust enforced at every stage.

DataGalaxy Implementation

With DataGalaxy as the system of record, every initiative at Sage flows through a structured process — from the first business request to live deployment and ongoing value tracking.

The DEM Lifecycle

Six stages shared between the Delivery Team, the DEM Team, and the Business Sponsor — with explicit handoffs and value checked at every gate.

Scaling Adoption Across the Enterprise

Scaling a data-product practice across a global business of millions of customers requires more than process — it requires trust by design. DataGalaxy enforces four non-negotiables at the metadata level — embedded into every asset and use case in the platform, not managed separately in policy documents. Regardless of domain or region, every product ships against the same standards:

• Security: Controls are designed in from the first ideation review, not retrofitted after launch.
• Privacy: Data minimisation and purpose limitation are enforced through metadata, not policy alone.
• Ethics: Use-case scoring includes ethical risk so that high-impact products are debated openly before delivery.
• Compliance: Regulatory alignment is a gating criterion at Business Readiness — not a final-mile review.

From Data Products to AI Products

Sage’s AI strategy did not begin with model selection — it began with the recognition that the same disciplines required for trustworthy data are amplified when AI enters the picture. Because DataGalaxy already held the lineage, ownership, and governance records for every data product, extending the DEM to AI products required no new infrastructure — only a new class of asset.

The same delivery teams, data stewards, business owners, and governance infrastructure that underpinned the data-product era were carried directly into the AI era — all operating within the same DataGalaxy environment. What changed was the scope of what they governed:

Each new AI initiative inherited trust, lineage, and governance by default — because all of that context already lived in DataGalaxy.

AI Governance

Most organisations treat AI governance as a compliance checkpoint — a gate to clear before shipping. Sage built something different. Their AI Governance, running on DataGalaxy, is a continuous, value-led process designed to accelerate the right AI initiatives while keeping risk visible at every step. Every AI initiative is linked in the platform to its underlying data products, governance rationale, and measured value — so nothing ships without a clear line of accountability, and nothing stalls without a reason. The process runs across six stages that mirror and reinforce the DEM:

• AI Request: Submission of the AI initiative for review.
• Governance Review: AI Governance validates triggers and routes to the correct review teams.
• Assurance & Approvals: All necessary approval teams review the request and advise on approval, conditions, or recommendations — with rationale captured.
• Design & Deliver: Approved initiatives flow into the same DEM-style delivery model used for data products.
• Value & Reuse: Progress is monitored; value and reuse of data and AI products is tracked continuously.
• Monitor Risk: Risk is monitored over time — decay is detected early and reuse is actively incentivised.

Value Lineage

Tying all of it together is value lineage — a living map within DataGalaxy that connects every report, data product, and AI product back to the business outcome it was built to serve. This is what turns the portfolio from a list of assets into an investment view: leaders can see what is being funded, what is being reused, and what is no longer earning its keep.

Outcomes

The combined discipline of the DEM, the data-product model, and AI Governance — all running on DataGalaxy — has produced measurable, business-visible results.

Key Takeaway

The Sage and DataGalaxy journey reinforces a quietly important truth: AI maturity is the dividend of data maturity. There is no shortcut — but there is leverage. When data is governed as a product, prioritised through a portfolio, and connected to outcomes through value lineage, every new AI initiative inherits trust by default.

For teams building toward the same destination, five implementation principles from Sage’s experience:

• Executive sponsorship is non-negotiable. A data and AI operating model survives only as long as leadership signals it matters. Without visible executive commitment, the model will be bypassed under pressure.
• Start with the process, not the technology. Tools change; the operating model is the durable asset. Build the framework first, then let technology serve it — not the other way around.
• Don’t underestimate change management. The hardest work is behavioural, not technical. Invest in enablement, celebrate visible wins, and make adoption a measured outcome in its own right.
• Treat data as a product from day one. Reusability cannot be retrofitted. Insist on ownership, lifecycle management, and business accountability from the first data asset you build.
• Measure value early and visibly. What gets measured gets reused; what gets reused gets funded. Make the value of every data and AI product visible to the leaders who control the budget.