Helping a manufacturer turn data governance from policy into something that actually works

The company was expanding cloud analytics and AI across finance, supply chain, sales and operations. Different teams used different definitions, important data had no clear owner, and getting access took too long.

Data and AI roles in a factory now pull together governance, data ownership, quality, master data, architecture, ERP and MES integration, analytics, AI governance and getting employees to use it all.

The first plan was to hire a Data Governance Director. A leader was needed. But one person could not set ownership, stand up metadata and quality tools, redesign access controls and drive adoption across every function.

What the company really needed was an operating model with leadership, architects, builders and domain stewards behind it.

We worked with the CDO, CIO and business leaders to pin down what governance had to deliver: trusted KPIs, clear ownership, faster access, cleaner master data, lineage, regulatory controls and support for AI.

The hiring plan split policy leadership from the people doing the work. It also made something clear: data stewards are business roles with real decision rights, not admin jobs handed out after the technical work is done.

Leadership:

• Chief Data Officer or Data Governance Director
• Owns the operating model, ownership, policy, adoption, value and executive alignment

Architects:

• Information Architect
• Data Governance Architect
• Data Security and Privacy Architect

Builders:

• Metadata and Catalog Engineers
• Data Quality Engineers
• MDM Engineers
• Governance Platform Administrators
• Security and Access Control Engineers
• Data Observability Engineers

Domain and adoption:

• Finance, supply chain, commercial and manufacturing data owners
• Business Data Stewards
• Data Literacy and Change Lead
• AI Governance Lead

We assessed people on:

• Data ownership and stewardship models
• Business glossaries, metadata, catalogs and lineage
• Data quality frameworks, controls and issue resolution
• Master data and critical data elements
• Role based access, privacy, retention and auditability
• Purview, Unity Catalog, Snowflake governance or similar tools
• ERP and manufacturing data understanding
• Getting people to adopt governance without turning it into red tape
• Proof that governance improved speed, trust or risk

We looked for governance people who could build and influence, not only write policy. The technical hires had to understand business ownership, and the leaders had to know enough about the platforms to make practical calls.

We also favored people who had rolled governance out step by step, through high value data products, rather than launching one big policy program with no connection to delivery.

Our advisors were governance leaders who have built working programs, not just written policy. They vetted the leaders on operating model and adoption, and the builders on metadata, quality, master data and access control. We sourced through our data governance and master data communities, where practitioners are known for programs that actually stuck.

A governance team tied to business results and to the technology work. Roles, decision rights and technical ownership were clearer across the central data team and the operating functions.

Trusted data products could be reused across analytics and AI. The company leaned less on a few experts and had a stronger base for scale, compliance and responsible AI.