Skip to the yard
Waxsealai

Four Challenges to Scaling Trusted Data Access Across the Enterprise

Shipped
Shipper
Franck Carassus
Readers
516

Trusted data sharing at scale depends on clear definitions, business ownership, easy discovery and access controls that work for people and AI.

Four Challenges to Scaling Trusted Data Access Across the Enterprise
Colleagues discuss data adoption
Colleagues discuss data adoption

Sharing data across an organisation can improve efficiency, collaboration and productivity, while helping companies get more value from existing data technology. The challenge is making that information useful at scale: employees and AI systems need to find data, understand what it means and use it under appropriate controls. Four connected problems shape whether that happens: trust, creation, consumption and access.

Build trust around a smaller set of data

Wider data access changes how an organisation works. In a siloed model, departments keep information within their own boundaries, and business users often rely on technical teams to prepare reports. Sharing data more broadly gives people greater scope to work directly with it, but it also exposes differences in how teams describe the same concepts.

“Active customer,” for example, may have different meanings across departments. A business glossary can establish shared definitions, reducing disputes and helping employees interpret data consistently. That semantic layer also gives AI systems context they don’t possess on their own. Without it, a tool may retrieve information that matches a term but misses the organisation’s intended meaning.

Clear definitions help, but users also need confidence that a dataset is accurate, relevant and dependable. Making every dataset available can increase uncertainty: people may not know which version to choose, who maintains it or whether it will remain current. A more useful starting point is a curated collection of business-ready data products. Each should have a named owner, quality standards and a data contract that sets expectations for its structure, freshness and reliability. Service level agreements can make the support commitment explicit.

This approach gives users a smaller number of assets they can learn to trust. It also makes the responsibilities behind each asset visible, so questions about quality or changes have a clear destination.

Make data creation a business responsibility

Collecting and cataloguing information isn’t enough; organisations need to turn it into assets that address real work. Central data teams can’t design every product for every department as demand grows. Responsibility for creating and maintaining data products should be distributed among data owners across the business, with the people closest to the data working alongside the people who need it.

That shift calls for a practical rollout. Start with business areas that already use data effectively and have clear problems to solve. A sales team, for instance, can work with data owners to shape products around its needs. The organisation can then observe how people use them, gather feedback and apply what it learns as the programme expands.

A shared self-service marketplace can support this work by giving teams a common place to publish products, document them and hear from users. It serves both as a route to discovery and as a collaboration space for improving data assets over time. Training and cultural change matter too: employees may be unfamiliar with finding and evaluating data themselves, so organisations need to help them build those habits.

Make discovery fit the way people work

Data that nobody uses creates no value. To encourage adoption, employees need to locate relevant information and understand how it applies to their tasks. AI agents need the same foundations: discoverable assets, clear descriptions and context about what the data represents.

Self-service data marketplaces aim to make that experience straightforward, with a familiar catalogue where users can search, read documentation, check reviews, contact an owner and see how colleagues use a product. Natural language search and AI agents can make discovery more accessible by letting people describe what they need in ordinary terms. Gartner rates deploying self-service data marketplaces as “essential” to scaling data consumption.

Ease of use still depends on the quality of what sits behind the interface. Search can point users to an asset, but it can’t resolve unclear definitions or make an unreliable dataset trustworthy. The glossary, ownership and quality commitments need to travel with the product so users can judge whether it fits their purpose.

Protect access while keeping it useful

Opening data to more employees and AI systems requires governance that protects sensitive information without blocking legitimate work. Access should reflect a person’s role and needs, while policies must also account for automated agents that can retrieve and combine data at scale.

Data lineage is a key part of that picture. It can show which human or AI system accessed an asset, how it was combined with other data and which decisions it informed. These records help organisations understand how information moves through workflows and support review of its use.

Marketplaces connected to corporate single sign-on systems can apply granular access controls, manage requests for permission and maintain an audit trail through lineage records. That gives users a consistent way to discover data while keeping access decisions and use traceable.

Scaling data adoption therefore takes more than publishing a catalogue. Organisations need shared language, accountable business owners, products that meet clear quality commitments and controls that keep use visible. With those pieces in place, a marketplace can help people and AI systems find and use data with less friction, while giving teams a practical way to improve the products they depend on.

Source: Franck Carassus · www.information-age.com

Same vesselRelated cargo