Causal AI is redefining decision intelligence, offering deeper insights and more precise enterprise decision-making than ever before.

In theCUBE Research’s latest analysis, discussed in the “Next Frontiers of AI” podcast, Stuart Frost, chief executive officer and founder of Gemino Software, joined theCUBE Research’s Scott Hebner, the podcast’s host. The conversation centered on the growing importance of causal AI and its potential to reshape enterprise decision-making.

Enterprises struggle to leverage AI for critical decisions. Although generative AI and large language models offer remarkable productivity boosts, their reliance on correlation rather than causation severely limits their effectiveness in business-critical applications, according to Frost. Current AI models can’t understand why events occur, hindering their capacity to provide genuinely insightful decision-making support.

“What our customers want is not another AI technology; they want to make better decisions,” Frost said. “Once you start going down that path of causal AI as opposed to correlational AI, you never go back. It’s just better.”

Why causal AI is changing the game

Businesses have long grappled with the limitations inherent in correlational AI technologies, which primarily predict outcomes based on historical data patterns without explaining their underlying causes, according to Frost. These traditional approaches can highlight problems, such as increased downtime in industrial operations, but they don’t explain the root cause or how to fix it. The shift toward causal AI addresses these shortcomings by offering insights into why events occur and what specific actions can mitigate or capitalize on them.

“Let’s say I’m in a refinery managing a process,” he explained. “Classic AI might indicate that downtime is increasing, but it doesn’t explain why or what to do about it. With causal AI, we develop a model that understands how a change in a variable — such as increasing the temperature — affects other variables in the system, like the pressure in the pipes. This approach enables us to identify root causes and precisely model what interventions to implement.”

This shift from correlation to causation allows businesses to identify issues, predict outcomes, and proactively influence results, enhancing their decision intelligence. Causal models let organizations simulate different scenarios, predicting the impact of each decision before committing resources, according to Frost.

“Once you’ve gone down this path and you’ve seen the power of it, you don’t go back as a data scientist,” he said. “It’s just better, fundamentally.”

The rise of causal AI is particularly significant in an era where enterprise decision-making demands explainability, transparency and accuracy. As regulatory environments and stakeholder expectations evolve, the need for decision intelligence becomes even more critical, according to Frost. Without explainable AI, enterprises risk making decisions based on unreliable or misunderstood information, potentially resulting in costly errors and loss of stakeholder trust, according to Frost.

“The fundamental issue with LLMs is they are just big correlational engines,” Frost explained. “They don’t understand the causal effects. And what we started to do is take the first steps to addressing that.”

Hebner further highlighted this critical distinction between correlation and causation, emphasizing its necessity for genuine decision intelligence.

“Correlation doesn’t necessarily imply causation and a prediction is not a judgment,” Hebner said. “You need judgments to make decisions.”

Making AI smarter: The next evolution of decision intelligence

Businesses today recognize the limitations of traditional generative AI solutions, and there is an increasing demand for AI capable of causal reasoning. Integrating causal knowledge graphs into enterprise AI infrastructures offers a compelling solution. These knowledge graphs enhance AI models by mapping cause-and-effect relationships, which in turn strengthens decision intelligence and improves decision-making, according to Frost.

“We use the knowledge graph then to drive more causal models at scale; that helps us make better decisions, and we can feed that causal experience and the actual practice of working on data and making decisions back into the causal knowledge graph,” Frost said. “Once that gets rolling, it becomes really powerful.”

The potential applications of causal knowledge graphs are immense, ranging from industrial operations to broader organizational decision-making processes. By capturing the dynamic elements of business knowledge, causal AI systems become far more effective and trustworthy, significantly improving decision intelligence across entire organizations, according to Frost.

“Agents and so on are not going to work very well if they don’t know what our business is about and what the constraints are and what the cause-effect chains are,” he said. “They can’t just pick it up from one small bit of context. They’ve got to understand how a business operates in order to help us.”

The rise of causal AI is also making advanced analytics more accessible, bringing sophisticated decision-making tools to organizations of all sizes, according to Frost. Instead of relying on the massive, brute-force computing infrastructures typical of generative AI, enterprises can now leverage streamlined causal models that use smaller, high-quality datasets to generate more precise, actionable insights.

As enterprises look for more reliable AI-driven insights, decision intelligence is emerging as a critical tool for shaping the future of intelligent automation. For a deeper dive into Hebner and Frost’s discussion, part of the  “Next Frontiers of AI” podcast series, check out their full conversation:

Image: SiliconANGLE/DALL-E