RCS World Tour | Paris

The Agentic Advantage: Pogoseat at RCS World by Google

Reflections from our panel at RCS World by Google in Paris — "The agentic advantage: smarter personalization, stronger relationships."

Our CEO, Kiran Patel, joined Fabian Hielscher of Google, Andreas Tussing of Charles AI and Guillaume Escolier of Wax on stage in Paris to talk about what agentic AI actually looks like when it has to do real work — not demos, not roadmaps, but live transactions with real money and real fans on the other end.

Two of the questions came my way: how the interface and the AI converge in RCS, and how agents handle genuinely complex workflows. Both go to the heart of what we've built at Pogoseat, so here's the longer version of what I said.

The model decides what to show. RCS decides how.

When people imagine an AI agent, they picture a model generating everything — the words, the buttons, the whole experience. Half of that is true. Fans text us in their own words — "can I get something closer to the ice, under $150?" — and the model reads it, replies, and carries the conversation. That side is genuinely open-ended.

But the moment it's time to choose or pay, the model steps back: it runs the conversation, but it never builds the interface you transact through. Free text is a terrible way to pick a seat, so the model's job is to read intent and decide what should happen next. What the fan actually sees at that point is native RCS: a carousel of real seats, a rich card with a real price, a "Buy" chip. Those elements come from fixed templates we control, populated from live inventory. And when the fan taps, that tap comes back as a structured action, not a sentence we have to parse and hope we understood.

Every rich element is one fewer place to hallucinate, because the options on screen come from the ticketing system, not the model's imagination. That separation is the whole trick:

The AI makes it feel like a conversation. The infrastructure makes it behave like a checkout. Keeping those separate is why it works.

Conversations are messy. That's the point.

The second question was about complex, multi-step workflows — and this is where messaging earns its place as a commerce channel rather than a marketing one.

Fans don't arrive with clean, single-step requests. It's gameday and the text says "I can't make it — can I return my tickets?" Behind that one message, the agent authenticates the fan, pulls the order, checks the club's return policy, releases the seats back to inventory, processes the refund and confirms — all inside the thread, end to end. The reverse case is just as common: "three friends are coming, can we sit together?" — find four contiguous seats, handle the upgrade difference, reissue the tickets.

It goes beyond ticketing operations, too. One of my favourite moments we've run: a fan scans into the venue, and a sponsor triggers a surprise-and-delight reward in the same thread. That's a brand-affinity moment no banner ad will ever buy — because it arrived in a conversation the fan was already having.

The payoff for clubs and venues is simple: fewer abandoned fans and more revenue per seat, precisely on the days when the box office is slammed and no human can take the call.

The mistake to avoid

If there was one warning I wanted the room to leave with, it's this: the biggest mistake brands will make is shipping agents that can talk but can't transact. Personalization that doesn't close the loop is just targeting with better manners. The agentic advantage — the thing that actually builds stronger relationships — is finishing the job in the thread: seat, pay, confirm.

That's what we've deployed with clubs and venues across multiple countries. RCS gives that experience a verified, rich, native home on the phone — and with Google pushing the channel forward, the thread is fast becoming the destination.

‍

Lear more about Pogoseat's Agent: https://www.pogoseat.com/news/agentic-ai-smarter-personalization-stronger-relationships