Zendesk AI speaks Babbel’s language

Learning a new language can be daunting. Babbel, a subscription-based app, makes it easier by helping users master new languages on the go and at their own pace.

“We focus on real-time conversations and teach useful vocabulary that can be put into practice in everyday life,” explains Gaia Brogini, Head of Customer Service at Babbel. “Both our website and app offer interactive lessons in 14 different languages.”

Babbel offers a mix of listening, speaking, reading, and writing exercises. With more than 25 million subscriptions sold, Brogini and her team receive questions 24/7 from users around the globe.

“Most of our agents are multilingual,” says Brogini. “And we always provide support in all the display languages offered by Babbel.”

But with support tickets pouring in from all over the world, the Babbel customer support team needed a better way to manage the volume of inquiries without sacrificing Babbel’s commitment to personalized support.

AI agents to the rescue

Babbel has been a Zendesk customer since 2008. When support team leaders saw the potential of AI agents to lighten their team’s workload—they were all in.

“When we started, we analyzed the types of tickets that we got—and targeted the ones that caused us the most friction, along with the types of tickets that came in a high volume,” explains Chris Boyd, Principal Tooling and Automation Manager at Babbel. “With the AI agents, we now automate requests from customers about things like their subscription status, how to manage their subscription, or changing their preferred language.”

After seeing how well the AI agents work, Boyd and his team are consistently adding new automations.

“We opened API endpoints to allow users to manage their subscriptions completely by themselves, integrating Zendesk AI with our backend systems,” Boyd says. “This frees up our support agents to focus on customers with more complex issues.”

Additionally Zendesk’s AI agents automatically tackle routine tickets across messaging, email, and chat channels. Subscribers get instant answers to simple questions like, “When does my subscription renew?”—no human agent required.

“That was one of our biggest pain points—and the AI agents have really, really helped,” adds Boyd. “Because of that, we can provide faster and more consistent support for all of our learners.”

No added headcount during peak periods

Babbel’s sales and associated support demand are highly seasonal. The company sees big subscription spikes in January, September, and November, coinciding with New Year’s resolutions, back-to-school, and holiday shopping.

During these key months, the volume of monthly support tickets can jump dramatically, often even doubling towards six-digit figures—all handled by a support team of about 50 agents.

“Whenever we’re trying to operate with the same amount of agents—it’s obviously quite hard to keep those peaks flat,” says Boyd. “We want to automate simple things so that our agents are free to reply to the complex queries far more quickly.”

AI agents let users get instant answers, smoothing out workload spikes.

“Thanks to Zendesk, we can maintain team size throughout the year,” Brogini adds. “And despite higher volumes, we still meet our KPIs.”

Zendesk AI uses a set of carefully designed “intents” to determine which tickets the AI agent can resolve independently and which require human intervention. For example, questions involving data protection requests need personalized support. This balance ensures customers receive both efficient and thoughtful service.

The results Babbel is seeing from implementing Zendesk AI speak volumes. Now, more than 50 percent of Babbel’s tickets are automated through AI agents. Email ticket volume, which used to comprise more than 65 percent of incoming inquiries, has fallen dramatically. Today, more than 45 percent of inbound requests are via the messaging channel.

Enhancing agent productivity with Zendesk Copilot

AI doesn’t just improve customer experience—it also helps Babbel’s human support agents work smarter and more efficiently. With Zendesk’s Copilot tool, AI-generated message summaries let agents quickly grasp the context of lengthy conversations, saving valuable time when ticket volumes are high.

“Suggested macros and enhanced writing tools are game changers—especially for agents responding in languages that aren’t their native tongue,” notes Brogini. “It helps ensure that replies are clear, professional, and personalized without adding extra effort.”

With AI agents handling the simpler tickets, Babbel’s agents spend more time on meaningful support.

“As a result of AI, agent satisfaction is higher,” says Boyd. “And they even actively contribute feedback to help refine the AI—becoming real partners in the process.”

Powering intelligent triage, multilingual support, and skills-based routing

Serving a global audience that speaks dozens of languages demands more than fast routing—it requires a deep understanding of every incoming request. And Babbel’s experienced, multilingual human agents get a boost from Zendesk Copilot’s intelligent triage capabilities.

“Copilot helps us to be more efficient by automatically matching tickets to agents who speak the customer’s language,” says Brogini. “This ensures a better, more natural support experience.”

Zendesk Copilot analyzes tickets as they arrive to understand intent, language, and complexity before taking action. Using AI-driven reporting and trend analysis, Babbel first gained clarity into what customers were contacting support about and in which languages. Copilot then consistently classifies tickets by intent and language, making it easier to route each request to the right expert from the start.

“Automatic language routing combined with skill-based ticket assignment ensures customers connect quickly with the right expert,” says Boyd. “It’s been essential in maintaining high-quality support while managing increasing volumes and complexity.”

Beyond language detection, Copilot also incorporates regional nuances—like local payment methods and subscription details—into responses, enabling highly tailored support at scale.

That intelligent triage and classification layer has also transformed how Babbel onboards new agents. Because tickets are clearly categorized and routed by skill, new hires only receive the issues they are trained to handle. As a result, onboarding time has dropped from three weeks to just one—a 67 percent efficiency gain.

“We’ve been able to introduce new agents much quicker because we now have the ability to just send them tickets on issues they’ve been trained on,” explains Boyd. “So we can get new agents handling actual support tickets within one week.”

Scaling for the future with more AI

Babbel is piloting Zendesk auto assist with a small group of agents, gathering feedback to optimize performance before wider rollout.

“We want to expand AI-powered tools like Zendesk auto assist,” says Brogini. “That way, we can scale up procedures that help agents respond even faster and more effectively.”

With an assist from Zendesk AI, Babbel’s support team is confident in its ability to continue delivering high-quality, personalized service to language learners across the globe.

“Zendesk AI hasn’t just made our support faster—it’s transformed how we work,” says Boyd. “It lets us focus on meaningful, personalized help while scaling to meet growing demand.”