Thomson Reuters CPO: Most Companies Are Wrong About AI – It's All About the 4 Pillars of Non-Agentic AI

In a world captivated by the dizzying promise of artificial general intelligence and fully autonomous agents, Kristy Grube, Chief Product Officer at Thomson Reuters, offers a refreshing and pragmatic dose of reality. Her message is clear: most companies are fundamentally misinterpreting where the immediate and most impactful value of AI truly lies.

Instead of chasing futuristic AI dreams, Grube champions a strategic focus on what she calls "non-agentic AI," underpinned by four foundational pillars.

Grube's perspective isn't about downplaying AI's revolutionary potential. Rather, it's a critical call to action for businesses to shift their gaze from the distant horizon of self-aware machines to the tangible, immediate opportunities that AI presents today – opportunities that augment human capability rather than replacing it entirely.

She argues that the fascination with "agentic AI" – systems capable of operating with little to no human intervention – often distracts from the pressing need to build robust, practical AI foundations.

Thomson Reuters, a global leader in information services, has invested heavily in integrating AI across its vast product portfolio.

Their strategy, as articulated by Grube, revolves around enhancing the expertise of legal, tax, and news professionals, not making them obsolete. This approach defines non-agentic AI: systems that assist, inform, analyze, and automate routine tasks, thereby freeing up human experts to focus on higher-value, complex decision-making and creative problem-solving.

To successfully implement this non-agentic AI strategy, Grube emphasizes the absolute necessity of mastering what she terms the "4 Pillars": Data, Process, People, and Technology.

Neglecting any one of these pillars, she warns, will inevitably lead to AI initiatives that falter or fail to deliver on their promise.

The first pillar, Data, is the bedrock of any effective AI system. Grube stresses that proprietary, well-governed, and high-quality data is an invaluable asset.

In an era where large language models (LLMs) are readily available, what truly differentiates an organization's AI capabilities is its unique data. Thomson Reuters, with its vast repositories of legal documents, financial data, and news archives, possesses a distinct advantage. However, merely having data isn't enough; it must be curated, cleaned, and made accessible in a structured way that AI can leverage effectively.

This involves robust data governance, ethical considerations, and ongoing quality assurance.

Next is Process. Even the most sophisticated AI is useless if it doesn't seamlessly integrate into existing workflows and operational processes. Grube highlights that AI should streamline and improve current ways of working, not create entirely new, cumbersome systems.

This requires a deep understanding of human-computer interaction and a commitment to redesigning processes with AI as an intrinsic part, ensuring that humans remain "in the loop" where critical judgment is required. It's about optimizing the flow of information and tasks, making AI a natural extension of human effort.

The third pillar, and arguably one of the most critical, is People.

Grube firmly believes that successful AI adoption hinges on investing in the workforce. This isn't just about training data scientists; it's about upskilling and reskilling the entire organization to understand, interact with, and leverage AI tools. Cultivating an AI-fluent workforce means fostering a culture of continuous learning, encouraging experimentation, and empowering employees to identify new opportunities where AI can add value.

The focus shifts from fear of replacement to excitement about augmentation, ensuring human expertise remains central.

Finally, there's Technology. While last, it's by no means least, and often where companies mistakenly start. Grube advises a pragmatic approach to selecting the right technological tools and platforms, including LLMs and other AI models, that align directly with the business's strategic objectives and the other three pillars.

It's not about adopting the latest shiny gadget for its own sake, but about choosing solutions that genuinely serve the data, fit the processes, and empower the people. This involves careful evaluation of vendor capabilities, scalability, security, and integration potential.

Grube’s insight offers a vital course correction for businesses grappling with their AI strategies.

By focusing on non-agentic AI and meticulously building out the "4 Pillars" – Data, Process, People, and Technology – companies can move beyond the hype and unlock the practical, transformative power of artificial intelligence today. It's a journey not about waiting for the future, but about strategically building it, one foundational pillar at a time.