AI agents and LLMs in the event industry: 5 use cases and how to get ready

AI agents have entered our daily lives. We plan, search, write and shop with them, and many people already can’t live without them.

The event industry won't escape the shift, and the most immediate promise is operational: impressive gains in time and productivity across event operations and analytics. LLMs could also change the codes of e-commerce itself, with a direct impact on ticketing and event commerce at large.

💡 Key takeaways

• Ticketing operations are shifting to natural language. What took a team days will take one person a few hours.
• Reporting is the sleeper hit: an LLM plugged into your ticketing, ads and CRM data does in minutes the weekly analysis that took days of exports.
• AI assistants are becoming a real acquisition channel. Your events must appear there, structured, before your competitors.
• AI support only works if the rights are defined: what the agent does alone, what needs validation, what stays irreversible.
• Buying is moving into the conversation. Your checkout must be able to live everywhere, not just on a web page.
• None of this exists without a solid foundation: open API, clean documentation, MCP server, headless ticketing backbone.

Why now?

Agentic AI is not a 2030 scenario. It's already in the numbers.

According to Adobe, AI traffic to US e-commerce sites grew 393% year over year in Q1 2026 and converted 42% better than non-AI traffic. 39% of American consumers already use AI to shop online. McKinsey sees agentic commerce orchestrating $3,000 to $5,000 billion in global revenue by 2030. And Semrush projects LLM-driven visits will overtake traditional organic search by 2028.

"LLM-driven visits are projected to overtake traditional organic search by 2028."

Ticketing, marketing, support, operations: nobody in the event industry escapes this shift. At Billy, we take it seriously. Here are the 5 use cases we see emerging, and what we think of each.

First, what's an AI agent?

An LLM (the technology behind ChatGPT or Claude) understands and generates language. A chatbot answers questions. An AI agent takes actions: it searches, compares, books and executes with the tools you give it access to. That last part changes everything for our industry.

01. Operations: “Deploy a tour” in natural language ⚙️

The agent executes directly in the ticketing system.

Our take: Interfaces won't disappear, but a growing share of operations will run through natural language. What took a team days will take one person a few hours.

The impact: On-sale setup compresses from days to minutes, and one ops person can run several tours in parallel. The bottleneck moves from execution to specification: pricing rules, quotas, presale logic.

Today or in 18 months? Today. It's already possible with modern platforms like Billy.

02. Data & steering: from the dashboard that observes to the agent that acts 📊

Yesterday, a dashboard told you what happened. Tomorrow, the agent will recommend and execute, under your control.

Our take: AI doesn't replace BI, it adds a layer: observation ⇒ diagnosis ⇒ recommendation ⇒ supervised execution. Agents can analyze data at scale to raise alerts and support decisions.

The impact: the weekly reporting ritual (exports from ticketing, ads and CRM, hours of manipulation) becomes a question you ask; revenue management shifts to continuous micro-adjustments. What was reserved for arena promoters with a data team becomes accessible to a producer running three shows.

Today or in 18 months? In between. Alerts and diagnosis work today with platform like Billy. Supervised execution will arrive in the months to come.

03. Augmented support: simple requests to the agent, the rest to humans 🆘

Ticket resend, name change, access question, resale possibilities: an agent handles the simple requests by applying your rules. Everything else is escalated to a human, with the context attached.

Our take: The challenge is not generating text. It's defining rights: what the agent does alone, what requires a validation, what stays irreversible.

The impact: Instant answers 24/7 on most of the volume, even during on-sale peaks. Your support people become escalation specialists focused on the cases that need judgment. Cost per request drops while the quality of the human queue goes up.

Today or in 18 months? Today. We’re already seing Agents handle triage, simple requests and generate detailed analytics. We still need to keep a human in the loop for advanced scenarios.

04. Discoverability: your events, found in the conversation 👀

Search is moving from "concert Paris September" to a complete request. The assistant compares, recommends and checks availability, all in one conversation.

Our take: this is a full acquisition channel, like Google or Instagram. Your events must appear there in a structured way, before your competitors.

The impact? Marketing teams will optimize for answers, not just keywords. Expect an GEO (Generative Engine Optimization) line next to SEO in your plan, and a new metric: your share of AI recommendations. Organizers with machine-readable catalogs will capture demand their competitors never even see.

Today or in 18 months? Today. Fans already ask their assistant for a plan. Ticketmaster launched a ChatGPT plugin to help fans access catalog (only available in the US for now).

05. Agentic commerce: buying without forms 🤖

Fans will no longer fill out a form. They’ll buy in the conversation, on your site or directly inside their assistant. The purchase funnel becomes a dialogue.

Our take: this only works with headless ticketing. The checkout must be able to live everywhere, not just on a web page.

The impact: the upsell logic changes. Insurance, merch, hospitality and experiences become one natural question in a dialogue instead of a checkout step. Revenue management extends into assistants, and your funnel analytics will need to track conversations, not just pages.

Today or in 18 months? Still emerging. The payment and trust rails are being built right now. ChatGPT launched its Agentic Commerce Protocol built with Stripe and Shopify. We’re seing this in the US for now.

Let's be honest about the limits

Agents make mistakes. They can hallucinate a price, misread a refund policy, or trigger an action you didn't want. Consequences can be dramatic. The good news is agents are improving. Every months, they keep getting better and better.

So the real design question is not "can AI do it" but "what is it allowed to do": act alone, ask for validation, or stay away because it's irreversible. Accountability stays with you, the organizer. An agent without boundaries can be dangerous so you need to define rules that your agent must follow to take actions. This can be done using skills, workflows and permissions.

What "agent-ready" actually means

None of these 5 use cases exists without a foundation:

• An open API: agents need programmatic access to your catalog and your checkout.
• Clean documentation: what an agent can't understand, it can't use.
• An MCP server: the standard that lets assistants like ChatGPT or Claude plug into your ticketing.
• Headless ticketing: checkout as a service, able to live in any interface.
• Solid data models: prices, categories, availability, conditions, readable in real time.

In other words: an agent-ready ticketing system is an open ticketing system with a solid foundation. Not one more feature.

Get ready: your 5-step checklist

1. Test your visibility. Ask ChatGPT or Claude about your next event. What comes back is your baseline.
2. Audit your ticketing stack. Open API? Public documentation? MCP server? Headless checkout? If it's no everywhere, that's the conversation to have with your provider.
3. Define agent rights. Write down what an agent may do alone, what needs validation, what is forbidden.
4. Connect your tools. Connect your ticketing platform to run your first prompt.
5. Start where it works today. Generate reports for advanced analysis, automate operations and support.

What about your team? Agents will not replace people. Roles shift from execution to supervision: defining rules, validating edge cases, steering agents. The ticketing manager of 2027 will spend less time clicking or exporting and more time deciding.

AI won't replace your ticketing tool

It just makes its openness non-negotiable.

The 5 use cases above won't arrive at the same speed, but they all point the same way: fans and teams will talk to your ticketing through agents. The winners will be the organizers whose events, data and checkout are ready to answer.

Some questions remain open. When good agents come to buy for real fans and scalper bots come to scrape your inventory, how do you welcome one and block the other? And when the sale happens inside an assistant, who owns the fan? Nobody in the industry has settled these yet. Exciting times to come!

In the meantime, if you want to see what agent-ready ticketing looks like in practice and meet Billy, talk to us :)