‘AI isn’t about replacing people’: Findability Sciences CEO Anand Mahurkar on enterprise AI

As companies race to deploy AI agents and generative AI tools, the bigger enterprise challenge may not be replacing people, but making organisations ready to use AI at scale, according to Anand Mahurkar, founder and CEO of Findability Sciences.

Speaking to ET Digital from his home in Boston, Mahurkar said the AI conversation is increasingly centred on tools such as ChatGPT, Claude, Codex and Gemini, while enterprises face a more complicated problem involving data, legacy systems and physical operations.

Agents are not the whole enterprise AI story

“Agents can now do most of the work” associated with knowledge workers, he said. But that does not mean enterprise productivity automatically rises.


In industries such as manufacturing, AI has to work across both digital systems and physical processes. Mahurkar cited a sugar mill, where trucks, weighing systems, laboratories, machinery, weather and traffic data all form part of the same operational environment.

Data comes before AI

Findability Sciences approaches this through what Mahurkar calls an ICUP framework: infrastructure, collection, unification, processing and presentation.

The company first conducts what it calls a “data census” to understand where an organisation’s data resides and how it can be unified.

“Unless you have unified data,” he said, enterprises cannot build the corporate memory needed for effective AI.

Not every problem needs an LLM

The company is “LLM agnostic”, using different commercial or open-source models depending on the use case.

Mahurkar said generative AI is only one part of the equation. Predictive AI can help identify machine failures, while computer vision can analyse images and translate them into operational decisions.

The approach has been applied across sectors, including manufacturing and bankruptcy services. Mahurkar cited Bourns, where Findability Sciences unified product information across formats and languages, and Stretto, where it worked with multiple live data sources and documents.

Manufacturing remains a big AI opportunity

Mahurkar sees manufacturing as an area where enterprise AI remains relatively underdeveloped. Applications, he said, can extend beyond employee productivity to reducing machine downtime, production losses, spoilage and resource waste.

For Indian businesses, he believes the question is no longer whether AI has value, but how organisations can implement it effectively.

“Your business is to make electronic components. AI is not your business,” he said, arguing that companies need the right processes, methodology and expertise rather than simply adding another AI tool.

From AI workshops to an AI factory

His broader argument is that enterprise AI needs to move from experimentation to industrial-scale deployment.

“The world of enterprise AI thinks workshops,” Mahurkar said. “You can’t develop at scale. You need a factory mindset.”

For enterprises moving beyond AI experimentation, the challenge may therefore be less about finding the next model or agent and more about building the data, infrastructure and processes that allow AI to work across the organisation.