From the moment you search on Google to the instant a chatbot replies in a call centre, artificial intelligence is quietly working behind the scenes of our digital lives. But while the promise of AI is compelling, turning that promise into actual value takes more than plugging in a chatbot.
At True Classic, a men’s fashion brand known for its well-fitted basics, Senior Customer Experience Manager Jordan Gesky has overseen a transformation that’s made AI a core part of customer service. The result is faster resolution rates, lower costs, and more satisfied customers. Jordan joined us on the VUX World podcast and shared how they did it and what others can learn from their approach.
Start with a simple problem
True Classic didn’t begin with a grand AI strategy. Their journey started in a place many companies can relate to: too many repetitive questions and not enough time. Customers asked about order tracking, returns, sizing – the kind of straightforward questions agents could answer in seconds, but that took up valuable time.
Their early automation was rule-based. If a customer typed ‘return’, they’d get a canned response with a link. It worked, but it felt impersonal. As Jordan says, customers didn’t feel “heard” or “secure”. So the team looked for a more natural, conversational solution. That led them to upgrade to a more advanced AI tool and eventually to generative AI, technology that could understand questions in context and respond in a way that felt more human.
Let AI do the easy work
Once they brought generative AI into their helpdesk, the impact was clear. By training the AI using past customer interactions, FAQs and business rules, they enabled it to handle a large portion of support tickets independently. Crucially, this freed up their human agents to focus on the complex issues, the ones where empathy, discretion or creative thinking are still required.
The numbers back this up as True Classic cut resolution times by 30%. And customer satisfaction didn’t drop, it increased. Why? Because customers were getting answers faster and agents had more time to help when it really mattered.
Give your AI a personality
One of the most overlooked aspects of customer-facing AI can be personality. True Classic understood that customers want an experience that aligns with the brand. So they gave their AI assistant a name, Crue, after their bestselling crewneck shirts and a voice that matched their tone: relaxed, witty and friendly.
It was about building trust. Customers are more willing to interact with an assistant that feels intentional and human. It makes the experience feel less like shouting into a void and more like having a real conversation.
Don’t set and forget
One of the clearest messages from Jordan’s experience is that AI is not a fire-and-forget solution. When they first went live with generative AI, she reviewed conversations, corrected errors and refined the assistant’s behaviour. Even now, the team regularly audits transcripts, checks customer feedback and updates the AI with new product information or policy changes.
After forgetting to update the AI when their return window changed from 30 to 100 days, Jordan learned this the hard way. It’s a small example, but it shows how quickly things can go off track if AI isn’t continuously monitored and improved.
Be strategic when using AI across channels
True Classic now uses AI across chat, email and even over the phone. But they didn’t launch everything at once. They started with chat and then added email support through the same system. Later, when they saw a surge in phone enquiries, they introduced a separate voice AI solution to handle that volume.
Each channel has different needs. Voice AI, for example, still sounds a bit robotic and requires extra care – pauses, intonation and real-time processing. But True Classic found it valuable for high-volume use cases like order tracking, especially since the system can hand off to email or text when needed.
Measure what matters
Jordan and her team look beyond resolution time or deflection rates. They also look at customer satisfaction, review transcripts, and conduct intent-based analyses to see when things go wrong, even if the customer didn’t leave feedback.
Manual reviews are still essential, but they’re also looking to layer in analytics tools to measure conversations at scale. As volume grows, the ability to measure effectiveness without reading every interaction becomes critical.
Be brave, be patient
One of the biggest surprises for True Classic was how well customers took to AI. Despite initial worries about customer acceptance, AI adoption was seamless. AI now handles many support tickets across chat, email and phone and customers are happy with the experience.
But it didn’t happen overnight. The secret was Jordan’s willingness to get in, test, tweak and improve. It was about committing to a long-term strategy of learning and refinement.
