Prashant K Dhingra on Agentic AI and Enterprise Value

In a time when almost every boardroom is discussing artificial intelligence, the real challenge has shifted from access to understanding what actually creates value.  

That is what made our conversation on The Executive Outlook with Prashant K Dhingra, Founder and Chief AI and Data Officer at DataKnobs, so meaningful. His perspective does not come from theory alone. It comes from years of working across some of the most influential organizations in technology and finance, including Microsoft, Google and JP Morgan. Yet what stands out most is not only his experience. It is the way he thinks. 

Prashant does not speak about AI as a shiny new tool that will magically solve everything. He speaks about it as something much more serious and useful. In his view, AI only matters when it improves decisions, strengthens workflows and helps businesses operate more effectively. 

That shift in thinking changes the whole conversation. 

The Moment the Enterprise Reality Became Clear 

As Prashant reflected on his journey, one turning point stood clearly. He had seen how earlier technology companies built products through software code. Then he watched the next wave emerge, where companies like Google have powerful AI-driven experiences on top of enormous volumes of consumer data from platforms like Search and YouTube. 

But enterprise environments, he realized, would always be different. 

Consumer platforms could be built at scale using freely generated public behavior. Enterprises could not. Their data was owned, protected, fragmented and deeply tied to internal business processes. That meant enterprise AI could never follow the same path as consumer AI. 

That realization became a defining insight for him. 

It was not enough to ask what AI could do in theory. The real question was how AI products would be built in environments where data was not open, where workflows were complex and where outcomes had to justify investment. That is where Prashant began to see the future more clearly. Enterprise AI would have to be grounded in business reality from day one. 

Why Agentic AI Became the Focus 

Long before generative AI became a mainstream headline, Prashant had already seen signs of how powerful AI could become. At Google, he worked on a scenario involving papers written by doctors, where AI could help automatically edit content based on prior patterns. This mattered because medical papers often took a very long time to review, partly because the right experts were difficult to find. Later, at JP Morgan, he explored how generative AI could be applied in finance. 

Across these experiences, a larger pattern started to become visible. 

AI was no longer just about speeding up simple tasks. It was becoming capable of supporting more complex kinds of work. That is what drew his attention toward agentic AI, AI and data products and the broader enterprise opportunities they could unlock. 

What makes his view especially practical is the framework he uses to think about value. He sees three major dimensions where AI creates real impact: 

Automation 

Personalization 

Creativity and reasoning 

This is not a vague vision of the future. It is a clear way for leaders to think about where to apply AI in the real world. 

Automation matters where scale is difficult for humans to manage alone. Personalization matters where broad segmentation is no longer enough. Creativity and reasoning matter where businesses need better ways to solve problems, generate ideas, or make more intelligent judgments. 

That is the lens through which Prashant sees the next wave of enterprise value. 

AI Is Not the Product. It Is the Transformation Layer 

One of the strongest ideas from the conversation was also one of the simplest. 

In a typical enterprise, AI is not a product. It is the transformation layer. 

That single idea changes how leaders should think about AI investments. 

Many companies still begin with the model. They ask which tool is most advanced, which interface looks most impressive, or which system is getting the most attention in the market. Prashant approaches it differently. He begins with the raw signal. 

Every enterprise already has raw data flowing through the business. It may come from customers, internal systems, markets, devices, operations, or transactions. But raw data on its own is often too noisy, too fragmented, or too weak to drive confident decisions. The role of AI is to transform that weak signal into a stronger one. Once that happens, the result becomes useful enough to improve a workflow, support a decision, or make a process better. 

That is where productivity starts to rise. That is where ROI becomes visible. And that is where enterprise value actually begins. 

This idea also explains why so many AI experiments fail to move beyond excitement. If the technology does not improve the signal inside the business, then it may still look impressive, but it will not create lasting value. 

Start With the Outcome, Not the Model 

There was one part of the conversation that felt especially important because of how practical it was. 

When asked what companies should do from day one while building AI-powered data products, Prashant did not begin with a tool recommendation. He began with discipline. 

His advice was simple: first identify the decision or workflow that needs improvement. Then determine what raw data already exists. After that, define what stronger signal would help improve that workflow or decision. Only then should a company decide which AI model is best suited to produce that signal. 

This order matters. 

Too many organizations start with technology because it feels exciting and immediate. But when AI enters the conversation before the business problem is properly defined, teams often end up building solutions that look advanced but remain disconnected from outcomes. 

Prashant’s framework does the opposite. It forces clarity. It helps teams focus on why they are building something before they decide how to build it. In enterprise environments, that is often the difference between a useful system and a wasted experiment. 

What Real Data Products Look Like 

To make the conversation more concrete, Prashant shared examples that showed how agentic AI and data products work in practice. 

One of the simplest examples he gave was LinkedIn. On the surface, it may look like a platform where people upload profiles and connect with one another. But underneath, it is also a data product. Profiles, recommendations, skills and connections all become signals. Those signals grow richer as the network grows. Recruiters can use them to assess talent. Marketers can use them to identify leads. Professionals can use them to explore opportunities. What begins as scattered user input eventually becomes a useful intelligence layer for many different outcomes. 

He then shared an example built by his own team: a stock assistant. This system takes multiple raw signals from the stock market and uses AI to produce a company’s momentum score and performance score. It also looks at stock options data to determine whether a certain price may be attractive for entering a position, while also showing demand and supply behavior around that price. Work that may take analysts many days can be made available much faster and in a more actionable form. 

But the deeper value is not just speed. 

The deeper value is that a user is no longer staring at disconnected market data. They are receiving a stronger signal that helps them make better decisions. 

Another example came from a large data center company. By analyzing voltage and current data from devices, Prashant and his team built a data product that could indicate device health, estimate remaining life and support depreciation decisions. Raw electrical readings may not look like strategic intelligence at first. But once transformed properly, they become deeply relevant to operational and financial decision-making. 

That is the real promise of agentic AI in the enterprise. It does not create value simply by sounding advanced. It creates value by making weak signals useful. 

How to Separate What Is Real From What Is Hype 

This is where the conversation became especially useful for business leaders trying to cut through the noise. 

Prashant offered a simple way to distinguish genuine enterprise AI from hype. If the AI system is working on real data that is continuously flowing into the business, that is a sign of reality. If it is connected to external tools and supporting real actions or workflows, that is another strong sign. But if it is sitting on static data and producing attractive outputs without changing how the business actually operates, then it is much closer to a demo than to enterprise transformation. 

That is an important test. 

Many AI systems look impressive on the surface. Their interfaces are smooth. Their outputs are polished. Their demos are convincing. But the real question is not whether the output looks smart. The real question is whether the system is alive inside the business. 

Is it connected to live workflows? Is it drawing from real signals? Is it improving decisions and outcomes in a measurable way? 

If not, then what an enterprise may be seeing is potential, not performance. 

Model Selection Still Requires Judgment 

Even after the workflow is defined and the data is understood, another critical question remains: how do you choose the right model? 

Prashant was clear that this is never a one-step decision. 

Experience plays a major role because experienced practitioners understand model capabilities and can reduce unnecessary experimentation. At the same time, experimentation is still necessary. Some use cases require simple reasoning, while others demand more advanced reasoning. Some domains use language in highly specialized ways. A word like "capital", for example, can mean something very different in finance than it does in general conversation. The same applies to healthcare and other specialized industries. 

That means model selection is not only technical. It is contextual. 

Leaders need to think about the nature of the problem, the language of the domain, the expectations of the workflow and the practical benefit to the user. A model may work well alone, but if it doesn't improve the workflow, it's still the wrong choice. 

That is why Prashant views model selection as a balance between capability, domain fit, experimentation and real-world usefulness. 

Where the Biggest Opportunities Are Emerging 

Looking ahead, Prashant sees the biggest opportunities where his three dimensions come together: automation, personalization and creativity. 

In automation, he pointed to high-volume environments like call centers in banking and insurance. In the past, reviewing every customer interaction would have required enormous human effort. Today, AI can monitor thousands of calls, identify where customers are facing problems, detect where agents may need training, highlight possible complaints and even surface signs of regulatory risk. This is not only about saving time. It is about improving service, oversight, and compliance at scale. 

In personalization, he described opportunities such as personalized drug development and advanced financial planning. These are powerful examples because they move beyond broad customer segmentation and toward more individual, precise support. Instead of putting people into large buckets and offering one recommendation to everyone, enterprises can begin serving individual needs in a much more intelligent way. 

In creativity, he described work around design generation, including product design such as bed sheet patterns and pointed toward future applications involving kitchens, furniture and even buildings. These examples expand the conversation beyond the back office. They show that agentic AI is also becoming relevant in spaces where design, imagination and reasoning play a major role. 

And then there are areas like drug discovery, which bring all of these dimensions together. Personalization, automation and reasoning do not sit separately there. They work together. That is what makes such use cases so powerful and so important to watch. 

The Bigger Shift Is a Leadership Shift 

What stayed with me most from this conversation was not just the technology. It was the mindset behind it. 

Prashant does not view the future as a competition to integrate AI into every aspect of the business. He frames it as a discipline. A discipline of identifying the right problem, finding the right signal and applying the right transformation to improve a real outcome. 

That is a more mature way to think about enterprise AI. 

For years, many organizations have operated by grouping people into broad segments and offering similar recommendations across large categories. But the future he describes is more precise than that. With the right data products and the right transformation layer, businesses can begin serving people more individually and more intelligently. Health needs can be understood more accurately. Financial planning can become more continuous and personalized. Operational systems can become more adaptive and more useful. 

In that sense, the enterprise becomes better not because it has more AI, but because it has better judgment embedded into the way it works. 

Final Thoughts 

What makes Prashant’s perspective stand out is the balance in it. He clearly sees how transformative agentic AI can be. But he is equally clear that enterprise value will not come from hype, novelty, or experimentation without direction. 

It will come from something more disciplined. 

It will come by turning raw signals into decision-ready intelligence. It will come by connecting that intelligence to real workflows. AI will be used as a transformation layer to make the business smarter, more responsive and more useful in critical moments, not as a spectacle. 

That is where agentic AI stops being interesting and starts becoming valuable. 

AI alone doesn’t create value. The right data strategy does. 

If you’re looking to build systems that truly work, connect with Complere Infosystem for expert guidance. And explore more insights from global leaders on The Executive Outlook.