$GRPN PEOPLE SERIES — #2: GIOVANNI LAGASIO The first person in this series, Adi Rajkumar, was about local marketplace execution. Giovanni Lagasio is about something different: How do you take complex global operations and rebuild them around AI, product thinking and operational rigor — without losing the customer in the process? Groupon introduced Giovanni as SVP, Global Operations. And his background makes the mandate interesting. Giovanni isn’t just an AI strategist. He is an operator working at the intersection of AI, product and global operations. Groupon describes him as an executive with more than a decade of experience transforming large-scale fintech, e-commerce and technology organizations, building intelligent systems that improve growth, resilience and operational excellence. Before Groupon, he worked across: Klarna Uber Just Eat Google At Klarna, Giovanni led global Product and Operations teams across: • Fraud • AML • Banking • B2C • B2B His work spanned multiple countries and an organization of 5,000+ employees, with a focus on modernizing service delivery, improving compliance and accelerating innovation. Before Klarna, Groupon says he led multi-market operational transformations at Uber, Just Eat and Google, producing measurable improvements in efficiency, revenue and customer satisfaction. That background matters because Groupon’s challenge is not simply: “How do we use AI?” The harder question is: How do you embed AI into the operating system of a global marketplace so that it changes how the company actually works? At Groupon, Giovanni leads the global operations agenda. His mandate includes: • building greater operational rigor • optimizing end-to-end processes • harmonizing operations across regions and channels • strengthening AI capabilities across operations and customer experience And unlike many corporate AI stories, we already have something measurable. Customer support. Groupon says its AI assistants, knowledge base and decision trees are now built and managed in-house, rather than depending on external providers. And the change in automation has been significant: AI now resolves more than 80% of customer-support conversations. At the beginning of the year, that number was around 40%. One important caveat: Groupon does not define what “resolved” means in this metric. Does it mean the customer’s problem was actually solved — or simply that the conversation ended without escalation to a human? That distinction matters. It is the difference between true resolution and deflection, and it is the single metric I’d most like Groupon to define more clearly. Even with that caveat, the move from ~40% to 80%+ tells us something important: AI is now handling a materially larger share of Groupon’s support workflow in production, not just in pilots or demos. But support automation is only part of the story. Groupon also says AI now scans deals every day to identify potential problems before they spread — helping merchants improve and allowing Groupon to intervene earlier. That connects directly to one of Groupon’s biggest historical challenges: trust. The company’s own customer-experience material ends with a very simple objective: “a Groupon you can trust.” This is why I find Giovanni interesting. His background is not: “AI expert joins Groupon.” It is closer to: Klarna → complex global product & operations Uber / Just Eat → multi-market operating environments Google → technology at scale Groupon → AI + operations + customer experience Of course, that alone does not prove the turnaround. The real test is whether better operations eventually translate into: fewer customer problems → lower friction → stronger trust → better retention → higher purchase frequency The first parts of that chain are already becoming measurable. Sources: Groupon leadership materials, Groupon LinkedIn, Groupon customer-experience materials.

$GRPN This is what Groupon means when it talks about becoming AI-native. A year ago, Groupon’s Data Platform Reliability team was spending most of its time on repetitive operational work — triaging alerts, digging through logs, investigating failures and following remediation playbooks. So they built an AI DataOps Agent using Google ADK. It doesn’t just diagnose problems. With guardrails, it can actually take action: • rerun failed jobs • scale clusters • trigger workflows • reset watermarks • clean staging data • escalate to humans when needed And now it’s operating at real scale. Over the last 90 days: 6,640 alerts 3,917 issues investigated ~59% coverage 11.2 min median first response 7.8/10 human-rated accuracy The most important part? Groupon didn’t reduce the operational scope. They changed who does the first mile of the work. Engineers are being moved away from repetitive remediation and toward higher-leverage work — including large-scale data platform migrations and product development. This is where the AI story at Groupon gets interesting: Not AI slides. Not buzzwords. Real automation → higher engineering productivity → lower operating friction → more capacity for growth. $GRPN