Last Saturday, we hosted AICon Chennai 2026 at IIT Madras Research Park (IITMRP) , bringing together AI practitioners, developers, architects, technology leaders, founders, students, and enthusiasts for a morning of learning, questioning, and exchanging ideas.
What started as an internal initiative through the Klizer AI Club is gradually becoming something bigger: a community where people can openly discuss not only what AI can do, but what it actually takes to make AI work in the real world.
We received 250+ applications, selected 125 participants, and welcomed 70+ attendees to the event.
But for me, the numbers were not the most important part.
What stood out was the curiosity in the room.
People were asking difficult questions. Speakers were sharing not just successes, but limitations and failures. Conversations went beyond prompts, chatbots, and the latest models into architecture, governance, cost, evaluation, technical debt, trust, and the role humans continue to play.
That is exactly the kind of AI conversation I believe we need more of.
Beyond the AI Demo
One theme came through repeatedly during AICon:
Building an impressive AI demo is becoming easier. Building an AI system that people trust and use in production is still hard.
Sivarajan N., Chief Technologist at TechnoTryst Web Solutions, spoke about moving from experimentation to enterprise adoption.
Drawing from more than three decades of enterprise engineering experience, he introduced the Enterprise AI Framework (EnAIF) and challenged us to think beyond models and demos.
One statement stayed with me:
“Autonomy is a cost, not a feature.”
Every additional level of autonomy introduces questions around governance, identity, security, reliability, and accountability.
The lesson for leaders is simple: we should not measure AI progress by how autonomous a system appears. We should measure whether it solves a meaningful problem and whether people can trust it.
What If AI Didn’t Need the Cloud?
PHILEMON P made this idea very tangible.
During his session on Edge AI, he switched off Wi-Fi completely and demonstrated AI running locally—including real-time speech-to-Tamil translation and vision detection on a Raspberry Pi.
No cloud dependency.
It was a powerful reminder that the future of AI architecture may not simply be about sending every request to increasingly powerful cloud models.
For many use cases, local-first with cloud fallback can offer advantages around latency, privacy, reliability, and cost.
As leaders and architects, we need to start asking a different question:
Where should intelligence live?
The answer will increasingly depend on the problem we are solving rather than which model happens to be most popular.
Can We Trust the AI Leaderboard?
Bharat Kulkarni took us into another important area: AI evaluation.
His session explored how models behave when they are being evaluated and how seemingly small changes to an evaluation harness can significantly change benchmark results—even without changing the underlying model.
This raises an important challenge for companies adopting AI.
Public benchmarks can help us understand models, but they cannot replace evaluating AI against our own business context, workflows, data, users, and expected outcomes.
Instead of simply asking:
“Which model is #1?”
Organizations should increasingly ask:
“Which model performs reliably for the problem we are trying to solve?”
Continuous evaluation in production will become just as important as model selection itself.
The Human Side of Agentic AI
Our panel discussion, moderated by Yassar Saleh , brought together Sai Mohan Vuchuru , Bharat Kulkarni , and PHILEMON P to discuss some of the less comfortable questions around AI engineering.
Where should we use agents, and where is human judgment still essential?
Does AI-generated code accelerate development today while quietly creating tomorrow’s technical debt?
How do leaders use AI to gain operational visibility without turning that visibility into micromanagement?
And perhaps most importantly, what can we learn from AI experiments that simply did not work in production?
I particularly enjoyed this part of the event because these questions rarely have simple answers.
We need more spaces where engineers and leaders can talk openly about what failed—not just showcase polished demos.
AI maturity isn’t about having all the answers. It is about asking better questions, experimenting responsibly, measuring what matters, and learning quickly.
From an Internal Club to an Open Community
AICon also marked an important milestone for us.
We officially launched the Klizer AI Club Community Portal at aiclub.klizer.com.
Klizer AI Club originally started as an internal space for our teams to experiment, learn, share ideas, and challenge each other’s thinking around AI.
Now, we want to open that conversation further.
Our goal is not to create another place filled with AI news.
We want to build a community where practitioners can share what they are building, discuss what is working and what isn’t, learn from one another, and meet people who are solving similar problems.
AI is moving too quickly for any individual, team, or company to learn everything alone.
Communities matter.
Thank You
A big thank you to everyone who joined us at AICon Chennai 2026 by Klizer AI Club at IIT Madras Research Park.
It was wonderful to see so many people come together to learn, exchange ideas, ask questions, and explore what is next in AI.
A special thank you to our speakers and panelists — Sivarajan N., Yassar Saleh, Philemon Philip, Bharat Kulkarni, and Sai Mohan Vuchuru — for sharing their experiences and perspectives so openly.
Thank you also to every attendee, volunteer, and member of the Klizer team working behind the scenes. Events like this only happen because many people quietly take ownership of dozens of details before anyone walks into the room.
And a very special thank you to Aswathy Mohan for hosting the event and bringing everything together so wonderfully.
The energy, conversations, questions, and curiosity in the room made the day special.
This is only the beginning for the Klizer AI Club community.
Explore AI. Exchange Ideas. Shape What’s Next.
Discover more from Reflections by Mohan Natarajan
Subscribe to get the latest posts sent to your email.