How Causal AI is Unlocking the Secrets Behind ‘Why’—And the Companies Already on Top of It

This shift could be groundbreaking, particularly in robotics and reinforcement learning, where AI agents would evolve beyond passive observation to active experimentation, leading to more reliable and efficient systems.

In a recent podcast, Robert Osazuwa Ness, Senior Researcher at Microsoft, delved into the evolving field of Causal AI, emphasizing its transformative potential. According to Ness, traditional AI—though highly effective in recognizing patterns and making predictions—often misses the mark when it comes to understanding the underlying reasons behind observed phenomena. As Ness explained, causal AI fills this gap by focusing on cause-and-effect relationships, a crucial distinction that could revolutionize industries reliant on deep analysis and robust decision-making.

Ness noted that causal AI builds upon Bayesian statistics, incorporating causal assumptions about how data is generated. This allows systems to not only interpret data correlations but to understand the mechanics behind them. The importance of this is profound in fields like systems biology, where causal AI helps to map complex molecular interactions and reveal biological mechanisms that standard AI models might overlook. Ness also discussed how causal graphs can provide a deeper understanding of biological systems, pushing beyond mere correlations to identify the actual causes of diseases, which is vital for fields like drug discovery and healthcare.

One of the standout features of causal AI that Ness emphasized is its ability to perform counterfactual reasoning. This means exploring “what if” scenarios—analyzing how changes in one variable could lead to different outcomes. This capability is particularly valuable in domains such as A/B testing in marketing or optimizing enterprise decision-making, where businesses constantly need to explore alternatives to make informed choices. By understanding not just what happened but why it happened, causal AI can help mitigate the risk of spurious correlations that could otherwise lead to biased or incorrect decisions.

Ness also pointed out that one of the most significant milestones for causal AI would be enabling intelligent systems to intervene in their environment. In his vision, these systems would not only observe data but actively gather interventional data—testing hypotheses in real time and learning from the outcomes. This shift could be groundbreaking, particularly in robotics and reinforcement learning, where AI agents would evolve beyond passive observation to active experimentation, leading to more reliable and efficient systems.

The Growing Importance of Causal AI in Modern Data Science

The applications of causal AI extend far beyond the theoretical. Ness highlighted that many questions in data science are inherently causal. For instance, when companies run A/B tests or assess marketing strategies, they’re not just interested in what worked but why it worked. Understanding the root cause of success or failure is key to replicating or improving those outcomes. Causal AI provides the tools needed to answer these questions more effectively, ensuring that decisions are based on a deeper, more reliable understanding of data.

Ness also tackled one of the major limitations of Generative AI and Large Language Models (LLMs)—while they can process vast amounts of information quickly, they often fall short when it comes to reasoning and critical thinking. These models, which mimic fast, intuitive decision-making (often referred to as System 1 thinking), struggle with the more analytical, slower thought processes required for complex problem-solving (System 2 thinking). Causal AI, by providing this analytical depth, complements generative AI, offering businesses a way to bridge the gap between fast insights and deeper understanding.

Companies Leveraging Causal AI

Several companies are already demonstrating the power of causal AI across different sectors, showing how this technology can be applied to solve real-world problems.

Causely, a platform founded by Ellen Rubin and Shmuel Kliger, is harnessing causal AI to automate the troubleshooting and management of cloud applications. Particularly focused on Kubernetes environments, Causely’s technology aims to eliminate human intervention in detecting and addressing operational issues, using causal AI to predict and resolve problems before they escalate. This not only improves the efficiency of cloud management but also significantly reduces downtime, which can have massive cost implications for businesses.

In the healthcare domain, Aitia (formerly GNS Healthcare), founded by Colin Hill and Iya Khalil, has been a leader in applying causal AI to drug discovery. Aitia uses multi-omic patient data and causal learning to uncover hidden biological mechanisms, which helps accelerate the development of drugs for diseases like Alzheimer’s and various cancers. Their use of digital twins—virtual simulations of biological systems—allows them to explore how different therapies might affect disease progression, thereby speeding up the discovery of breakthrough treatments.

Another key player in the healthcare space is Insitro, founded by Daphne Koller, which combines machine learning with biology to predict effective therapies for complex diseases. Their approach is particularly focused on diseases of the liver and central nervous system, where traditional drug discovery methods have struggled. By integrating causal AI into their models, Insitro is able to predict which treatments are most likely to succeed, reducing the time and cost associated with bringing new drugs to market.

Howso, co-founded by Dr. Michael Capps, Mike Resnick, Chris Hazard, and Gaurav Rao, also stands out in the causal AI landscape. Howso’s platform goes beyond standard AI models, offering advanced tools for exploring cause-and-effect relationships in data. Their platform provides businesses with capabilities such as prescriptive analysis, counterfactual reasoning, and what-if scenarios, empowering organizations to make better, data-driven decisions. Howso’s emphasis on ethical AI and robustness also ensures that their models are both transparent and reliable, helping companies navigate complex decision-making processes with greater confidence.

Why Causal AI is Gaining Traction

The rise of causal AI can be attributed to several key factors. First, as Ness mentioned, the limitations of current AI models—particularly LLMs—are becoming more apparent, especially in critical fields like healthcare, finance, and enterprise decision-making. While these models excel in generating insights quickly, they often lack the transparency, interpretability, and reasoning capabilities needed for more complex tasks. Causal AI addresses these shortcomings by offering a more methodical and explainable approach to AI.

Second, businesses are increasingly hesitant to trust AI systems that operate as “black boxes,” especially when it comes to making high-stakes decisions. Hallucinations, bias, and the lack of interpretability in current models have raised concerns about their reliability. Causal AI, by explicitly modeling the relationships between variables, helps to reduce bias and improve the accuracy of predictions. This makes it a more appealing option for industries that require both speed and precision, such as healthcare, finance, and operations management.

Finally, the integration of causal AI with generative AI offers a powerful combination. While generative AI can process vast amounts of data quickly and provide intuitive insights, causal AI brings the analytical depth needed to fully understand the data. Together, these technologies promise to revolutionize enterprise decision-making, combining rapid analysis with robust, methodical insights that can guide more strategic decisions.

The Future of Causal AI

As causal AI continues to evolve, its applications are likely to expand across a variety of fields. From improving the efficiency of cloud infrastructure to accelerating drug discovery and enabling more ethical AI systems, causal AI is positioned to play a central role in the future of artificial intelligence. By enabling AI systems to understand not just correlations but the actual causes driving outcomes, causal AI provides the transparency and robustness needed to push AI into new, more reliable territories.

As Ness pointed out, one of the next major breakthroughs will likely be in embodied AI and robotics, where intelligent systems will not only observe but also interact with their environments, learning from interventions. This development could lead to smarter, more adaptive robots that can better navigate and understand the world around them, pushing AI further toward true autonomy.

📣 Want to advertise in AIM Research? Book here >

Picture of Anshika Mathews
Anshika Mathews
Anshika is the Senior Content Strategist for AIM Research. She holds a keen interest in technology and related policy-making and its impact on society. She can be reached at anshika.mathews@aimresearch.co
Subscribe to our Latest Insights
By clicking the “Continue” button, you are agreeing to the AIM Media Terms of Use and Privacy Policy.
Recognitions & Lists
Discover, Apply, and Contribute on Noteworthy Awards and Surveys from AIM
AIM Leaders Council
An invitation-only forum of senior executives in the Data Science and AI industry.
Stay Current with our In-Depth Insights
The Most Powerful Generative AI Conference for Enterprise Leaders and Startup Founders

Cypher 2024
21-22 Nov 2024, Santa Clara Convention Center, CA

25 July 2025 | 583 Park Avenue, New York
The Biggest Exclusive Gathering of CDOs & AI Leaders In United States
Our Latest Reports on AI Industry
Supercharge your top goals and objectives to reach new heights of success!