The key to AI value: Four ways to unlock returns with a robust data strategy

Facing mounting pressure to deliver measurable returns on artificial intelligence (AI) investments, organizations are urgently seeking to overcome the practical challenges that stand in the way of measurable value. Chief among these challenges are issues related to data curation, governance, and usability, which directly impact return on AI investments. This article lays out a rationale and roadmap for building the robust, semantically rich data foundation essential for scalable, high-impact AI initiatives that drive value.

Insight

KPMG 2026 US Technology Survey report

In today's rapidly evolving digital landscape, data and technology leaders face unprecedented opportunities and complex challenges.

Organizations overwhelmingly recognize the promise of AI as a revenue-generating innovation, with all respondents to the KPMG 2026 Technology Survey expressing confidence that AI will drive business value by 2026.¹ Yet, many companies are unsure how to achieve that value. For many, the crux of the issue lies not in navigating complex AI use cases, but in addressing the state of the underlying data that fuels AI capabilities.

Data readiness is emerging as the critical bridge between AI ambition and tangible return on investment (ROI). Traditional data architectures—often fragmented, siloed, or designed primarily for human consumption—cannot deliver the speed, accuracy, or scale that advanced AI requires. Having a strong data strategy involves not only migrating to the cloud or upgrading technology infrastructure, but also fundamentally rethinking how data is structured, integrated, and governed. Curating high-quality, semantically rich datasets allows AI models to leverage data for dynamic forecasting, intelligent automation, and personalized customer experiences.

In essence, data readiness is the foundation that transforms AI from a cost center into a strategic engine of growth.

While employee trust in AI output is steadily increasing — with 76 percent of our survey respondents expressing confidence in AI-generated recommendations¹— there is still progress to be made, particularly when it comes to mission-critical business functions. This requires a shift toward AI decisions that are consistently grounded in transparent, structured, and high-quality data.

Knowledge engineering plays a pivotal role in this process by creating a framework where every AI-driven insight can be traced and explained. By implementing detailed knowledge graphs, organizations help ensure that AI outputs are not only accurate but also explainable. The further organizations progress into AI adoption, the more requirements for data rigor and resilience increase. Data hardening — or the strengthening of data management processes — ensures that data products are robust, secure, and reliable. This involves refining practices for data quality, governance, and privacy controls, which collectively safeguard the integrity of information and support compliance. This level of transparency, as a critical part of a comprehensive governance strategy, allows stakeholders — including regulators, customers, and internal teams —to understand and verify the logic behind AI output, fostering greater confidence in AI-powered decisions.

Essentially, reliable data is the bedrock upon which trust in AI systems is built.

Insight

Why knowledge engineering is the key to AI agent value

Unlocking sustainable value from agentic artificial intelligence (AI) is a strategic imperative for your business. The discipline of knowledge engineering allows organizations to capture and structure their information for machine use, so they can scale their AI efforts enterprise wide.

Although optimism about AI remains high, only 31 percent of respondents in our tech survey said they are seeing measurable ROI from multiple AI initiatives.¹ This gap between the potential benefits and actual results often comes from overlooking the true operational costs of data products or not accurately measuring their value against clear business outcomes. Simply creating a data product because it’s technically possible isn’t enough—if it doesn’t support AI tools in generating meaningful business value, then it isn’t worth the investment.

To tackle these challenges, organizations should use a straightforward, repeatable process to guide decisions about data investments. By adopting a disciplined approach—including benchmarking—companies can make smarter choices about where to focus their efforts. AI can leverage knowledge graphs to quickly and precisely measure value, so organizations can estimate both gains and costs based on these calculations. This transparency makes it easier to track development timelines, resource needs, and strategic priorities, ensuring the original value proposition stays intact. Finally, using a data value chain provides a structured, quantitative way to assess and maximize value at every stage of the data lifecycle within an AI initiative.

In this way, data is central to deriving ROI from AI initiatives.

Our tech survey reveals that 92 percent of US organizations are already investing in the development of agentic AI,¹recognizing it as the cornerstone of the future hybrid workforce. As organizations seek to empower AI agents to reason, adapt, and handle increasingly complex and autonomous tasks, the necessity for interconnected data that is organized according to context, relationships, and meaning is undeniable. This involves building knowledge and context engineering engines that can classify certainty and uncertainty, optimize the balance between machine decision rights and human oversight, and establish guardrails that operate seamlessly within the flow of transactions rather than being applied retroactively.

Critical to the knowledge engineering discipline are semantic layers and ontologies. Semantic layers act as abstractions, organizing complex data into understandable concepts, while ontologies define relationships and meanings within the data. These enable AI systems to interpret information contextually, classify the reliability of insights, and deliver actionable recommendations that drive business growth. Moreover, knowledge processing engines not only enhance the interpretability of AI outputs but also provide mechanisms to distinguish between high-certainty and low-certainty information. This classification allows organizations to set appropriate decision boundaries, ensuring that AI agents have the autonomy to act on clear, well-defined scenarios, while humans remain involved in situations where ambiguity or risk is higher.

It is clear that, without context-rich data, the full promise of agentic AI will not be realized.

In the race to realize AI’s transformative power, it’s clear that the return on your investment depends on your data. No matter how advanced your AI ambitions or how visionary your organization, your ROI will likely fall short without a robust, modern data foundation. Success hinges on treating data not as an afterthought, but as a strategic asset—one that’s reliable, explainable, and designed for the agentic future.

Organizations that prioritize data readiness, including the integration of context and knowledge engineering practices, position themselves to harness AI’s full potential by ensuring every insight is transparent and actionable. By investing in your data infrastructure today, you lay the essential groundwork for the explosion of AI agents to deliver meaningful, trustworthy outcomes—unlocking sustainable business value and positioning your organization at the forefront of tomorrow’s AI-driven innovations. Ready to unlock the full value of AI? Start by investing in your data infrastructure today.

Footnote
¹KPMG Technology Survey, “From Automation to AI: Tech Leaders are focused on AI,” (Jan 2026)