In this episode of Software People Stories, Gayatri Kalyanaraman reconnects with Zafar, who brings nearly four and a half decades of experience across technology and multiple waves of transformation—from mainframes and personal computing to the internet, cloud and digital transformation.

His observation about AI is simple but profound: previous technology shifts gave humans more powerful tools; AI is beginning to democratize intelligence. And with AI agents, even execution is being democratized.

Zafar is enthusiastic about AI's potential, but cautious about the hype.

Giving thousands of employees access to AI tools is not, by itself, transformation. The real question is: what has changed in the way the organization creates value?

For him, meaningful enterprise value will come from redesigning workflows, integrating data, changing decision rights and building domain-specific AI systems—not simply adding AI assistants to existing processes.

AI also changes what leadership means.

When information, analysis and execution are increasingly available on demand, leaders cannot rely simply on having the answers. Their role shifts towards asking better questions, creating clarity, setting context, exercising judgment, aligning people and taking accountability.

The future may be less about managing people or technology and more about orchestrating human and AI capabilities around outcomes.

Perhaps the most thought-provoking part of the conversation is Zafar's concern about cognitive over-dependence.

We already outsource calculation to calculators and recall to search engines. AI takes this much further by allowing us to outsource drafting, analysis and even parts of reasoning.

His question is not whether we should use AI—we clearly should—but:

What mental capabilities do we deliberately continue to exercise ourselves?

He argues that productive struggle—the process of wrestling with a problem, making mistakes and changing our minds—is part of how we develop critical thinking and intellectual independence.

The conversation then moves from cognition to something even more fundamental: human relationships.

AI is infinitely patient, always available and highly personalized. But what happens if frictionless AI relationships begin replacing the messy, challenging and sometimes uncomfortable relationships we have with real people?

That raises a bigger question about human agency: are we eventually going to outsource not only memory and reasoning, but also judgment, empathy and emotional support?

Zafar also reflects on the societal implications of increasingly autonomous AI.

AI has extraordinary dual-use potential—it can accelerate scientific discovery and healthcare, while also being used for surveillance, manipulation, cybercrime and misinformation.

As AI becomes increasingly agentic, his message is clear: governance and human oversight need to advance alongside capability.

He argues for a risk-based approach to AI governance: a tool rewriting an email should not face the same level of scrutiny as an AI making decisions about healthcare, employment, credit or public security.

Ultimately, this is a conversation about much more than AI technology.

It is about what happens to organizations when intelligence becomes abundant, what happens to leaders when answers become cheap, and what happens to humans when machines become increasingly capable.

And perhaps the most important question we leave with is:

As AI becomes better at thinking for us, what must we continue learning to think for ourselves?

Zafar can be reached at https://www.linkedin.com/in/zafar-ahmad-4704b76/

My guest today is Paramu Kurumathur: IT Leader, Author. Here is the snippets of interesting stories around how AI has changed his life and how he’s impacting AI - From building AI in 1980 to using AI to code, write, research and imagine the future

What makes Paramu's story unusual is that his AI journey actually began in 1980, when his M.Tech project explored automatic program generation using structured data and if-then rules. Four decades later, he finds himself using generative AI in almost every facet of his life.

From preparing a draft of his family will to asking Gemini to explain a television show, AI has become a surprisingly useful personal companion for Paramu.

At work, he uses AI coding tools such as Cursor to turn ideas into working software—even when he doesn't know the programming language particularly well. One of his recent projects used AI, Python and a vector database to help people discover government schemes they may be eligible for.

For Paramu, this ability to take an idea and quickly turn it into something tangible is one of the biggest "aha moments" of AI.

AI has also found its way into Paramu's writing. He uses it for Sanskrit transliteration, translations, research, introductions and making his writing more conversational.

But he is equally conscious of AI's limitations. Hallucinations become particularly problematic when dealing with subjects where the available data is limited, such as astronomy and ancient Indian texts. This leads to an important observation about the need for good foundational data.

Perhaps the most interesting question in the conversation is what Paramu would never delegate to AI.

His answer: the core idea of his books, personal relationships and the human connections involved in coaching and consulting.

AI can help us create, research and accelerate—but there are parts of our lives where the relationship itself is the work.

When asked what is underrated about AI, Paramu points to its potential for social and collective benefit.

From identifying potholes and holding contractors accountable to helping people discover government benefits, he believes some of the most meaningful applications of AI could emerge outside traditional productivity use cases.

The conversation ends with a glimpse into the future—from AI agents that can behave in surprisingly autonomous ways to Paramu's wish that technology could help people communicate when they cannot express themselves easily, or even objectively understand pain.

And, of course, there is a rapid-fire round before we finish—with one final thought: perhaps five years from now, Paramu might return to the podcast… in AI form!

A conversation about AI, human judgment, social good and what we choose to keep uniquely human.

The conversation ends with a rapid-fire round.

Favorite AI use case?

A tool.

What is overrated about AI?

Nothing.

What is underrated?

AI for social and collective benefit.

What would you never delegate?

The core message of his work—and the human relationships that matter.

And perhaps the perfect closing thought: when Gayatri imagines that five years from now Paramu might be sitting next to her in AI form, Paramu signs off by saying he enjoyed the "stimulating" discussion—while joking that he always confuses stimulating with simulating.

A fitting end to a conversation about humans, machines and everything in between.

Paramu can be reached at https://www.linkedin.com/in/paramu-kurumathur-01489a6/

and paramukurumathur.com