While tech headlines continue to chase ever-larger AI models capable of passing bar exams and penning passable poetry, it looks like another kind of intelligence is quietly taking root.
Soft AI — defined as AI systems designed to perform specific, narrow tasks using approximate reasoning, pattern recognition, and flexible decision-making — seeks to mimic human-like thinking by embracing ambiguity.
But what does this mean for businesses today and why should anyone care?
Soft AI Explained
At the heart of soft AI is a principle that mirrors human cognition, based on the idea that the world isn’t binary. Humans don’t make decisions in ones and zeros. We make them in context, based on incomplete data, in the presence of doubt. Soft AI seeks to reflect this messiness.
According to Badr el Jundi, CEO of Greenverse Partners — a company applying AI to both industrial risk detection and healthcare diagnostics — “soft AI is less about brute computational force and more about adaptability, approximation and human-like reasoning.”
Jundi compared it not to agentic AI, which is task-driven and goal-pursuing, but to a more flexible kind of intelligence — a system that weighs uncertainty and navigates ambiguity much like the way people do. Rather than relying on massive foundational models, Greenverse’s systems integrate compact neural networks with components of fuzzy logic to interpret real-world signals in ways that allow for nuance and flexibility.
But what’s fuzzy logic anyway and why does it matter now?
Defining Fuzzy Logic
Fuzzy logic — a soft computing concept introduced in the 1960s for intelligence systems that aren’t rigid and can better adapt to real-world environments — is foundational to soft AI. Unlike traditional logic, which frames truth in black and white, fuzzy logic teaches machines to see the world the way we do — not as yes or no, but somewhere in between. For example, a system might say a signal is 70% safe and 30% risky, rather than only label it ”normal” or “abnormal.” That kind of nuance is what makes all the difference in the real world.
Still, fuzzy logic is only one part of Greenverse’s approach. “We combine fuzzy inference with neural learning architectures and adaptive signal models,” Jundi told me. “It mimics the way humans make decisions.
This becomes crucial in fields like industrial safety, where systems must interpret noisy signals and respond in real time — or in healthcare, where patterns in early-stage disease can be faint, inconsistent and easy to overlook.
According to Jundi, this is an area where Greenverse Partners shines. He noted that the company calibrates these models specifically for each domain. “In healthcare, we focus on physiological signal variations over time. In industrial use, it’s environmental pressure, flow, and vibration patterns,” Jundi explained. “The underlying platform is adaptable, but the data models are domain-specific,” he added.
Purpose-built AI Systems
The rise of soft AI is challenging the idea that larger AI models — which companies like OpenAI continue to push for — produce better and more accurate outcomes. But the reality, according to Jundi, is that scale doesn’t always bring clarity in risk-sensitive sectors like critical infrastructure or emergency care.
“In safety-critical environments, precision, explainability, and responsiveness often matter more than scale,” he said. “Smaller, purpose-built systems can outperform massive models by being tailored, interpretable and highly dependable.”
That sentiment is echoed in a 2024 study by Deloitte which found that 42% of organizations deploying AI in high-risk settings preferred smaller, domain-specific models over generalized ones — largely because they are easier to test, validate and explain.
Beyond Automation
Soft AI isn’t trying to replace human judgment. In fact, as Jundi noted, its power lies in how closely it can work with us.
“Ethical innovation means building systems that earn trust,” noted Jundi. “We’re not aiming to replace human judgment; we’re creating tools that augment and inform it.”
It’s a philosophy that resonates with AI ethicists, too. As Fei-Fei Li, computer scientist and co-director of the Stanford Institute for Human-Centered Artificial Intelligence, put it: “the future of artificial intelligence is not about man versus machine, but rather man with machine. Together, we can achieve unimaginable heights of innovation and progress.”
Business Implications
For business leaders looking to get the best out of AI, the takeaway is that the next wave of AI innovation won’t come solely from larger, more complex systems. It will come from intelligence that can adapt, interpret and collaborate more effectively.
For Jundi, soft AI offers a model for that future, enabling more intuitive decision-making, faster responses to change and a greater capacity to handle uncertainty. And it does all this without requiring the compute burden or energy demands of massive models — making it particularly suited for edge deployment.
“Organizations should focus on agility, interdisciplinary collaboration and cultivating a culture that embraces decision-making under uncertainty,” Jundi advised.
Already, there are signs of this transition. Edge AI — which often relies on lightweight, task-specific models — is growing significantly. Global spending on edge computing is projected to reach $378 billion by 2028, driven by demand for real-time analytics, automation, and enhanced customer experiences. Much of that growth is powered by systems designed to understand context locally — without needing to send every decision to the cloud.
The Future Is Adaptive
Looking ahead, the promise of soft AI isn’t just better outcomes. It’s about making the relationship between humans and machines work better — where humans collaborate with systems that don’t just compute, but adapt to the fast, unpredictable rhythms of the real world.
“Over the next five years, expect breakthroughs not just in model architecture, but in how AI integrates into the world around us,” Jundi said, adding that this will happen “through lightweight intelligence, edge computing and systems that understand not only data, but context, nuance and intent.”
In the end, soft AI might not be as loud as other AI systems like generative AI or agentic AI. But it may prove to be the most intuitive — and perhaps the most trustworthy — form of AI we’ve seen so far. For business leaders listening closely, that could make all the difference.
