An Olympic volunteer takes time off, travels, and signs up months ahead to be part of the Games once in their life. So when 130,000 apply, 18,000 make it through, and everyone wants to know when their shift starts, where to report, or whether their accreditation is sorted, a slow answer costs more than a support metric. It can leave someone outside a venue in the cold during a morning that can’t be repeated.
That was the problem in front of Leonardo Cohen, who led volunteer support systems for Milano Cortina 2026. He had run workforce systems for the FIFA World Cup in Qatar, so this was familiar territory, and it still gave him pause.
A record-setting Winter Games, two cities, 116 events, two languages, a deadline that doesn’t move, and nowhere near a call center’s worth of people to absorb it.
It was massive. The international side, the complexity, the volume coming in. Massive.
Leonardo Cohen, Workforce Systems Expert, Milano Cortina 2026
The challenge
The questions never stopped. Applications, role assignments, schedules, health and safety, accessibility, and then the live ones during the Games, when someone needs an answer in ten minutes, not tomorrow. In Italian and English, across a country, against a clock that doesn’t move.
Adding people wasn’t on the table. Whatever Leonardo’s team built had to clear the bulk of that volume by itself and pass only the cases that truly needed a person to the right person, the first time. Get the routing wrong at this scale and you’re not annoying a customer; you’re stranding a volunteer who showed up to help.
The Games would open on schedule. The only open question was whether anyone would answer.
The audit
Leonardo had one rule going in: everything in one place. A volunteer’s first application and their panicked question mid-shift had to live in the same system instead of being split across a chat widget, an inbox, and a ticketing tool that didn’t connect. At Games scale, every seam between tools is a place where a request can go missing.
He also wanted something his teams could stand up quickly and run themselves, in the languages volunteers were writing in. One system for automation, live agents, and email tickets, not three vendors stitched together. Text did all of it on one platform.
The solution
Text AI Agent met nearly every volunteer first. Across almost 49,000 conversations, it cleared the routine volume on its own, sorting Italian and non-Italian volunteers into the right flow without anyone managing it. After 448,000 answers, the queue still hadn’t piled up.
What it couldn’t resolve, it handed off cleanly. Sensitive or complicated cases became tracked tickets with a full history and a clear path to the right department. More than 72,000 email tickets were handled through the same system.
Email-based support is still the big thing.
Leonardo had braced for a difficult onboarding across two languages and dozens of volunteer roles. After a half-hour walkthrough, agents were live the same day.
The training was straightforward. Easy to get agents going on it.
The result
There were 130,000 applicants, 18,000 volunteers, and one system from the first application to the closing shift. Of the almost 49,000 conversations that came in, only about 650 ever reached a human. Text handled the rest.
The volunteers were happy with it. Leonardo built his own reporting with Text’s APIs and scored sentiment across 73,000 interactions: 67% positive and 9% negative. That is a strong result for a support operation under this kind of pressure.
It was the easiest and most complex API I have ever used. I was able to extract everything by myself from the system to analyze the data. Such a powerful connection. The data was absolute gold.
What Leonardo keeps coming back to isn’t a feature; it’s the partnership. He is already applying what he learned to what comes next: the World Cup, LA28, and the Games after that.
I've worked with the big vendors. Never had this level of support. The future here is supervised AI agents, smart ones, humane, can do small talk. That's where all of this goes.
Give every customer an answer when volume peaks
Milano Cortina 2026 used AI for the routine and people for the cases that needed judgment, resolving 99% of almost 49,000 conversations without an agent. See how ChatBot’s AI customer service agent can give your team dependable coverage when demand surges.
Start your free 14-day trial. No credit card required.
