During the panel I moderated at the Goa AI Tech Conclave 2026, I kept referring to ChatGPT. It is the AI tool I use, so the name came naturally. Siddharth, representing Google, kept gently bringing Gemini back into the conversation.

After the session, someone took me aside and put it in terms I could hardly misunderstand: saying ChatGPT with Google on the panel was rather like introducing me as OSD to Minister X (Minister X shall remain unnamed), when I work with Minister Rohan Khaunte.

Fair point. I suspect Gemini will come to mind a little more readily next time.

Our panel, AI Powered Tourism: Enhancing the Goa Experience, brought together Siddharth Prakash from Google, Milind Anvekar from Open Destinations, Jitendra Kanojia from IHCL and Krishnakanth Naik from IBM.

I wanted to connect their perspectives to a practical question: what would AI actually change for someone visiting Goa?

That led me to Palolem.

A thousand travellers, the same recommendation?

Behind my question was a concern. If recommendations keep reinforcing already popular places, we could end up using sophisticated technology to give everyone much the same holiday.

Siddharth pointed towards a different possibility. People are bringing more complex questions to Gemini, supplying richer context about the experiences they want.

There is a difference between asking for the best places in Goa and describing a trip with elderly parents, an interest in local food, a preference for quiet surroundings and two days to explore. Those details change what a useful answer should look like.

For me, that connects directly with Goa Beyond Beaches. A visitor interested in architecture, a family seeking village experiences and someone looking for time in nature should be able to discover different sides of the destination.

But a more personal answer still depends on good information. That became one of the most useful threads of the discussion.

What the audience brought into the conversation

Two concerns stood out to me from the audience poll: the quality and fragmentation of data, and how to bring local businesses into AI-enabled systems.

Milind reinforced these practical challenges in the Goa context. My takeaway was that we need to organise scattered information and make participation possible for the businesses delivering the experience.

Consider a small homestay, an independent guide or a village experience provider. An AI system can only do so much if the information it needs about their services is missing, inconsistent or out of date.

For me, bringing data together means creating a reliable, connected foundation. It does not necessarily require every organisation to surrender its information to one central database. Shared standards and appropriate access could also help systems work together.

Equally, asking small businesses to “adopt AI” is only the beginning. They need a practical way to participate: manageable tools, support with their information, and a clear reason to invest time in the process.

The poll was a snapshot of the room, not a statewide survey. Even so, it brought the conversation firmly back to implementation.

Dilip, the boatman, and the busy homestay owner

We also spoke about Dilip, the boatman. How could AI help someone like him get business? For me, that question made the discussion about local participation concrete. Could a visitor looking for a boat experience find his service, get reliable information and make an enquiry that actually reaches him?

This is where I see a practical role for AI: answering routine questions from information the owner has approved, gathering booking requirements and helping manage enquiries. Where it has reliable access to availability, it could assist with that too, bringing the owner in for exceptions or decisions that need their attention.

These were possibilities we explored, rather than claims that a particular system was already doing all of this for Goa’s businesses. The purpose, though, was clear to me: give the host more room to be a host.

The owner should be able to finish a conversation with the guest in front of them without feeling that every minute costs another booking. Dilip should have a better chance of reaching a customer while remaining the person who brings knowledge, character and care to the experience.

The work behind the holiday

Milind also brought attention to the operational work that travellers rarely see. My recollection is of a discussion about contracts, digital systems and the repetitive administration that technology can help teams handle.

A recommendation still needs to connect to something a business can deliver. Behind a seemingly simple holiday are products, rates, conditions and people making sure the details line up. Reducing repetitive work there could give teams more time for judgement and service.

Milind’s point about skills was practical: having AI available is one thing; equipping people to use it is another. Krishna then offered a different way of approaching that challenge.

Skills for people, skills for agents

As I understood Krishnakanth’s perspective, we may not need to train every business owner extensively in AI. A group of experts could instead build specialised, skilled AI agents that people can use for the tasks they need help with.

That shifted the question for me. Alongside asking how we skill people to use AI, we should ask how much expertise we can build into the tools themselves.

Think again of the homestay owner. Experts could build an agent around routine guest enquiries and booking requirements, with suitable information, permissions and clear points for involving the owner. The owner would use that capability without having to become an expert in building or directing AI systems.

My takeaway was that good design could reduce the learning burden. People would still need to understand what the tool can do, check important decisions and know when to step in. But the specialised work of shaping the agent could sit with the experts.

Jitendra widened the human side of that discussion. Anxiety about AI and jobs, he reminded us, reaches senior management too.

We often focus on students preparing for their first job. Experience, however, does not remove uncertainty about how work might change. Leaders have to keep learning alongside the teams they ask to adapt.

A seven-hour delay, and a question about trust

One question I put to Jitendra made the possibilities of connected data tangible. Drawing on the Tata connection between Air India and IHCL, I asked: if someone has endured a seven-hour delay on an Air India flight and then arrives at the Taj, would the hotel want to know?

Imagine the difference that knowledge could make. The hotel might anticipate a late arrival, simplify check-in or greet an exhausted guest with greater understanding.

A traveller might welcome the sharing of a flight delay while wanting to keep other details private. Who they are travelling with, for instance, may be something they have no wish to disclose across an entire network.

Permission to help with a delayed arrival needs a boundary. It should not quietly become permission to circulate every other detail of the journey.

An old idea from Cranfield

The conversation reminded me of a project I worked on years ago at Cranfield, for the Google Chrome team. There was a certain irony in returning to that memory while sharing a stage with someone from Google.

Our idea involved simple buttons or controls through which people could choose categories of information they were willing to share. Shopping data, for example, could be a deliberate choice.

In our imagined model, if a hundred people contributed shopping data and that pool generated revenue, a portion would flow back to them through an agreed formula.

It was a thought experiment, with plenty of unresolved questions about measuring value and distributing it fairly. It felt utopian then. I still find its underlying question worthwhile: what say should people have in the use of their data, and how might they share in the benefit?

For a traveller, the benefit could begin with a smoother arrival or a more relevant recommendation. The commercial question goes further. Both depend on a choice people can understand and make, including the choice to decline.

Could tourism learn something from UPI?

A holiday crosses many businesses and services, but the traveller experiences it as one journey. UPI offered a useful reference point for separate organisations working through a common framework.

I came away with the impression that there was openness to exploring cooperation where responsibilities and safeguards were clear. That was an opening for further discussion, rather than a formal commitment from the companies represented.

A tourism framework would need to address permission for specific purposes, access that expires when a task ends, appropriate separation of information and accountability when something goes wrong. Those requirements would need their own design; the UPI analogy alone does not resolve them.

AI sovereignty added another layer. Who governs the systems we depend on? Who decides how information can be used? How do local priorities influence those decisions?

For Goa, I believe those priorities must include visitors, businesses and the communities whose daily lives form part of the destination. The governance deserves as much attention as the technology.

Rediscovering the pleasure of a good question

There was a personal takeaway too: I remembered how much I enjoy moderating.

My work across government, the private sector and digital platforms has given me different ways of looking at the same problem. On a panel, those experiences become questions about what an idea means for a business, a policymaker, a visitor or the person expected to make it work.

The pleasure comes from listening closely enough to find the next question.

I left thinking about Palolem, Dilip’s boat, the homestay kitchen and the tired guest arriving after a seven-hour delay. Each brings the discussion back to people.

The reason to return

My strongest takeaway was about what we should protect as the technology improves.

I suspect AI will increasingly become an expected part of travel: quick answers, easier bookings and less repetitive administration. Those conveniences will matter, but as more destinations offer them, the human experience will matter all the more.

For me, the lasting reason to return is the connection with people. The host who makes time for you. The boatman whose stories stay with you. The person who notices that you have had a difficult day and responds with kindness.

That is what I want AI to make more space for. Let it handle the heavy lifting, while people bring the warmth, judgement and personality that make a place memorable.

About the panel

Roles below follow the event’s published speaker listings.

• Siddharth Prakash: Head of Product Strategy, Google.
• Krishnakanth Naik: Program Lead, IBM.
• Jitendra Kanojia: Vice President, Artificial Intelligence, Indian Hotels Company Limited (IHCL).
• Milind Anvekar: Senior Vice President, India Operations, Open Destinations.

Panel: AI Powered Tourism: Enhancing the Goa Experience, Goa AI Tech Conclave 2026.

This article reflects my recollections and takeaways from the discussion; panelists’ contributions are paraphrased.

Event speaker directory