The New Data Challenge: Serving Humans and AI at the Same Time

For years, enterprise data platforms have been built around a single audience: people.

Reports, dashboards, and analytics tools all exist to help humans access information, spot trends, and make better decisions — often at their own pace, filling in the gaps with judgement and context along the way. But as AI becomes embedded in everyday work, that assumption is changing. In Singapore, the government is aiming to equip 100,000 workers with practical AI skills by 2029, reflecting the growing expectation that AI will play a routine role in how organisations operate and make decisions.

AI has become a second consumer of enterprise data, and an unusually demanding one. It doesn’t check in occasionally or tolerate ambiguity; it needs trusted, contextual, and continuously accessible information across every system, database, and document an organisation holds. The question this raises isn’t a narrow technical one. It’s whether the architecture underpinning enterprise data was ever built to serve two such different audiences at once, and what has to change if it’s going to serve both well.

One Data Estate, Two Very Different Audiences

The gap becomes clearer once you compare what each audience actually needs.

• Humans access data periodically, work through curated dashboards and reports, and can apply judgement to fill in context or tolerate minor gaps.
• AI and agents require always-on access, depend on machine-readable structure and clear metadata, and cannot infer meaning that hasn’t been explicitly governed or labelled.

The implication is uncomfortable for a lot of organisations: an architecture designed purely around human consumption doesn’t just under-serve AI — it can actively work against it, producing systems that are technically “AI-enabled” but practically unreliable the moment real decisions depend on them.

Why Most Architectures Weren’t Built for This

AI adoption is already widespread: According to a survey by McKinsey, 88% of respondents report regular AI use in at least one business function. Momentum is also building in Singapore, where Microsoft announced a S$7-billion investment (US$5.5 billion) in the country’s cloud and AI infrastructure from 2025 through 2029. Yet most enterprise data environments were never designed with AI in mind.

While these environments evolved to support business operations, analytics, and reporting, data is often spread across ERP platforms, CRM systems, and a growing number of cloud applications. This fragmentation has long created operational and governance challenges, but AI is raising the stakes significantly with Gartner predicting that 90% of organisations will adopt a hybrid cloud approach by 2027.

A business analyst can navigate disconnected systems, reconcile conflicting reports, and apply judgement when information is incomplete. AI cannot. Its outputs are only as reliable as the data it can access and understand. When information is fragmented, inconsistently governed, or stripped of business context, AI inherits those limitations.

The situation becomes even more complex as AI-generated content, insights, and recommendations increasingly become part of the enterprise data estate themselves. Organisations are no longer simply managing data created by humans. They are beginning to manage data created by both humans and machines, creating new questions around trust, traceability, and accountability.

This is why AI readiness has become inseparable from data readiness. Before organisations can scale AI with confidence, they must first ensure that the foundations supporting it were built for the current era of data consumption.

What Has to Change in the Architecture Itself

Serving both human users and AI requires more than connecting another tool to existing data sources. It demands a fundamental shift in how organisations manage the flow, governance, and meaning of information across the enterprise.

1. Integrating Across Systems, Databases, and Documents

The first priority is creating a more connected data estate. AI generates the most value when it can draw from information across business functions rather than isolated silos.

That means bringing together structured data from business systems alongside unstructured content such as policies, project documentation, and knowledge repositories. Without this level of integration, AI can only work from a partial view of the organisation.

2. Governing Access for a Non-Human User

Governance has traditionally focused on controlling what people can see and do. AI introduces a new challenge: enabling systems to access information while maintaining the same standards of security, compliance and accountability.

As organisations deploy more AI assistants and agents, governance must ensure that AI interactions and agentic workforces remain secure, auditable, and aligned with existing permissions and policies.

3. Adding Context and Meaning

Access alone is not enough. AI also needs to understand what information means within the context of the business.

Metadata, business glossaries, and lineage information help provide that context, allowing AI to recognise relationships, interpret business definitions, and generate responses that better reflect organisational reality.

4. Enabling AI to Reason and Support Action

The goal is not simply to help AI find data, but to help AI generate outputs that decision-makers can trust.

The most mature environments combine connected, governed data with business context and traceability. This provides the foundation for AI to deliver responses that are more accurate, explainable, and actionable — moving beyond information retrieval towards meaningful business support.

Where Qlik and Snowflake Fit in This Evolving Ecosystem

As organisations modernise their data foundations for AI, the focus is shifting from simply storing and analysing information to creating connected, governed and AI-ready data environments.

This is where platforms such as Qlik and Snowflake can play complementary roles. Qlik helps organisations integrate and prepare data from across disparate systems, while Snowflake provides a unified environment where structured, semi-structured, and unstructured data can be managed and analysed together.

Together, these capabilities support the broader shift: moving from fragmented information estates towards a connected and trusted data foundation. Regardless of how advanced AI becomes, its ability to deliver meaningful business value will ultimately depend on the quality, accessibility, and context of the data behind it.

Serving Both Humans and AI Starts with the Same Foundation

As AI becomes a new consumer of enterprise data, organisations face a challenge that extends beyond adopting new tools. They must create a data foundation that serves both human understanding and machine intelligence.

That means connecting information across systems, governing access consistently, and providing the context that enables AI to generate trustworthy outputs. Technologies such as Qlik and Snowflake can help support this shift, but the bigger priority is building a data architecture designed for both audiences from the outset. Because ultimately, the value of AI depends on the quality, accessibility and trustworthiness of the data behind it.