Bringing Intelligence to the Edge: Chetan Bansiwal on Real-Time AI and the Future of Automation
In conversation with Tarun Singh Rajput, CMO of Valuebound, Chetan Bansiwal—founder and AI innovator—dives deep into the evolution of AI-powered real-time analytics, his entrepreneurial journey, and the future of human-machine collaboration.
Back in 2019, a stolen car on a quiet New Delhi street became the catalyst for a technology company that would soon redefine the possibilities of real-time AI. Chetan Bansiwal, a serial entrepreneur and builder at heart, had just exited his first company and was looking for his next challenge. He was drawn to two domains: cybersecurity and artificial intelligence.
AI felt like the right frontier. It was accessible, fast-moving, and most importantly, open. “AI was something that I was quick—it was very quick to learn and the resources were available freely,” he said. YouTube, blogs, and forums offered an informal curriculum, and Bansiwal wasted no time diving in.
It was around the same time that a close friend’s car was stolen—despite the presence of security cameras and even police involvement. The vehicle was never recovered. “The cameras right now are not smart enough... they are very dumb sensors... they do not do any kind of inferencing,” Bansiwal noted. This real-world failure revealed a glaring gap: surveillance systems were passive collectors of data, not active protectors. What was needed was intelligence on the edge—real-time AI that could understand, infer, and alert.
From Dumb Sensors to Smart Safety
Thus began a new chapter. Bansiwal and his team started developing real-time analytics solutions that could interpret camera feeds at scale. They focused first on building high-accuracy models, then layered on specific analytics. “The main focus was to do realtime analytics over a camera feed and on scale,” he explained. This vision soon led to national-scale applications: projects on expressways and highways to detect vehicles, tunnel surveillance in hilly regions for natural disaster or militant threats, and even facial recognition-based security for major public events including Republic Day and Independence Day celebrations.
He also talked about one of their most impactful interventions: improving highway safety through smart alerts. On the Delhi-Meerut Expressway, a tragic collision involving a wrong-way school bus inspired Bansiwal's team to propose a network of Variable Messaging Signs (VMS) placed every 50 meters. These would be activated instantly when wrong-way driving was detected, allowing other drivers to switch lanes and avoid fatal accidents. “We don’t want crashes onto the road… and that is an easier save compared to destroying a tire,” he said, highlighting the difference between passive deterrence and proactive prevention.
AI's Next Evolution: From Developer Tool to Product Playground
Beyond computer vision, Bansiwal has kept a close eye on the rise of LLMs (Large Language Models). While many see them as technical tools, he believes the real transformation will happen when non-developers—particularly product thinkers—begin to harness these systems. “A lot of time for the developers will be freed… this is going to be a time for the product managers… to shine,” he said. These individuals can now outline a vision, and the AI can help build it.
This democratization of creation is shifting the power dynamic. He compares it to Instagram: users shine on the platform, but it’s the engineers behind the scenes who made that shine possible. In the same way, Bansiwal sees a split between power users who will leverage AI and the builders who create the systems they rely on.
However, accessibility remains a hurdle. While LLMs are great at indexing the internet and surfacing answers to highly specific queries, there's still a gap between building tools and delivering them globally. “The problem is still not the browser getting built—the problem still is that how will this browser actually reach that guy in the other part of the world,” he observed.
Specialized AI Agents Before Jarvis
While many dream of a universal assistant—an AI "Jarvis" that can handle anything—Bansiwal argues that we’re not quite there yet. People still struggle to grasp what an LLM can actually do, especially in the absence of clearly defined use cases.
“You are skeptical of its performance and… do not have those containerizations in your mind,” he pointed out. Instead, he suggests building specific AI agents: a virtual HR, a virtual sales assistant—models that align with existing business structures and responsibilities. “That is much more sustainable compared to giving a generic LLM to anybody.”
He envisions this phase lasting the next three to five years, after which a unified chat interface might finally emerge. But until then, clear boundaries and defined roles are essential for AI adoption.
The Human Touch Remains Irreplaceable
Even with rapid automation, Bansiwal doesn’t believe all domains can be handed over to AI. “Sales will never be automated,” he declared. “Anything that requires a human touch… will never be replaced.” He extends this to caregiving, healthcare, hospitality, and the arts. These sectors, he believes, will continue to rely on human intuition, empathy, and nuance—qualities no algorithm can yet replicate.
Building the Next Five Years
Looking ahead, Bansiwal is optimistic. As automation grows, he sees humanity entering an era of abundance—especially in terms of time. “It is a good thing to have a lot of time,” he said, especially for entrepreneurs who often lack it most. But this abundance must be paired with equitable access and distribution. He imagines a future where basic sustenance is a right, made possible by logistics, automation, and good intentions.
Ultimately, he sees his role not just in building technology, but in helping shape the rules by which AI systems will operate. “If we also write some rules in this era… five years down the line those rules will make us shine.”
Chetan Bansiwal’s journey from passive surveillance to proactive AI is more than a technical evolution—it’s a reflection of how technology, when paired with human intent, can save lives, shape industries, and perhaps even shift the very structure of how we live and work. In this critical moment, as LLMs and AI agents redefine what's possible, builders like him are not just creating tools—they're designing the future.
