Back
EP 008
26 Min
Rebag sells Chanel, Hermès, and Louis Vuitton, where one mistake can cost a customer's trust. Brandol Guerra, who leads customer service there, says AI can absolutely feel premium, but only if you teach it the attributes that are non-negotiable.
Most AI-in-support conversations are about volume and efficiency. Luxury is different: customers expect premium, personalized service, not a generic chatbot. Brandol Guerra is Associate Director of Customer Service at Rebag, the luxury handbag resale marketplace, after years at global BPOs like Sutherland and Teleperformance. On this episode of the Fini Podcast, he explained how to make AI feel premium, what he refuses to automate, and how a Six Sigma mindset sharpens both.
Meet Brandol Guerra
Brandol has spent over a decade in customer experience leadership, moving from large BPOs into an in-house role at Rebag, where the product is high-value and trust is everything. He brings a Six Sigma background focused on eliminating waste, and he is leading the integration of AI into a business that cannot trade quality for efficiency.
Making AI feel premium
For Brandol, premium starts by defining the attributes you want the AI to deliver, then training for them. An AI agent is an extension of your business, one more team member, so it has to recognize when someone is upset or frustrated and respond accordingly. That depends on getting the inputs right: the key questions for each scenario and the keywords the AI should identify. He frames training an AI almost like teaching it about human behavior, which is why you map everything before going live. Premium also means efficient, because every mistake is costly, so the goal is an agent that connects and resolves quickly, grounded in a strong knowledge base so accuracy is never the thing it has to worry about.
What stays human
Brandol's rule for what to automate is friction. AI is great for troubleshooting and how-to questions, but the moment a customer has invested too much time or the process carries too much friction, the AI should escalate to a person who can pick up where it left off. He is specific about what he will not automate: a customer ready to purchase (you never want to add friction to a sale), a broken payment or checkout, and change management, like explaining how the business is evolving after a launch. Those are moments where a human needs to take over to avoid losing the customer.
From BPO speed to in-house experience
Having run support at big BPOs, Brandol is candid about the difference in-house. In a BPO, every minute has a cost, so teams rush to end the call, which quietly sacrifices the attention to detail that defines real customer experience. In-house, you can build processes and teach your team the value behind every step, why empathy, why attention, why the flow. His Six Sigma lens is about eliminating waste, and his favorite example is an AI that flagged "this customer is frustrated" but not the details or the next step, so a top agent doing well over a hundred cases a day still wasted time interpreting it. The fix was not more automation, it was rewriting what the AI was prompted to capture, based on direct feedback from the agent using it.
Spotting promoters and detractors early
The move Brandol credits for lifting Google and Better Business Bureau ratings is reading the interaction early. His team is trained to tell within the first few minutes whether a customer is a promoter or a detractor, then adapt the flow instead of marching a frustrated customer through a script. The other half is proactive: after a tough interaction, the team calls the customer back, and about half the time people were upset not because the issue went unresolved, but because they could not express what they needed in the moment. Once they cool down, an agent helps them shape it. As he puts it, the difference between what a customer wants and what they actually need is where good resolutions come from, and it is not a sin to say "let me circle back."
What support leaders should take from this
Define the non-negotiable attributes first. Decide what premium means for you, then train the AI on the keywords and questions that detect it.
Automate by friction, not by category. Let AI handle troubleshooting and how-tos, and escalate the moment a customer has spent too long or the process gets heavy.
Never automate a sale or a broken checkout. A customer ready to buy, or stuck paying, needs a person, not a protocol.
Fix the input, not just the model. When AI output is unhelpful, the problem is usually what you asked it to capture. Get feedback from the agents using it.
Read promoters and detractors early. Adapt the flow in the first minutes, and call detractors back once they have cooled down.
Train teams on need versus want. Teach the team (and the AI) to find what the customer actually needs, not just what they first ask for.
Listen to the full episode
Brandol goes deeper on training teams to embrace AI, using Six Sigma to cut waste, and where to start in a luxury brand, in the full episode of the Fini Podcast. You can connect with him on LinkedIn.
AI that feels premium and hands off cleanly when trust is on the line is what Fini is built for. Book a demo to see it in action.
Leo: Welcome back to the Fini podcast. I'm your host Leo. Today we're talking about a challenge most people don't think about when they hear AI in customer service: how do you automate support in a luxury environment where customers expect premium, personalized service, not generic chatbot responses? My guest is Brandol Guerra, Associate Director of Customer Service at Rebag, the luxury handbag resale marketplace. Rebag deals in Chanel, Hermès, and Louis Vuitton, high-value goods where customer trust is everything and mistakes are costly. Brandol has spent over a decade in customer experience leadership, including years at global BPOs like Sutherland and Teleperformance before moving in-house to Rebag, where he's leading the integration of AI into a business that can't afford to sacrifice quality for efficiency. Brandol, welcome to the show.
Brandol Guerra: Thank you, Leo. Thank you for the space. There are a lot of topics for us to discuss today.
Leo: Most AI conversations focus on volume or efficiency, but luxury is different. Customers expect premium, personalized service. How do you deploy AI in a luxury environment without killing that experience?
Brandol Guerra: Let's define what people consider premium. As a customer myself, I'm picky, and from a premium experience I expect more than empathy and soft skills, it goes beyond that, and this is where AI takes place. The first thing is defining the attributes you want the AI to deliver. There may be chatbots, AI voice, and agents you need to train, and it's about how you train them. An AI agent is an extension of your business, one more agent, and it should be trained to deal with people, to recognize if someone is upset or frustrated. To ensure that works, you need your key questions in place, what you expect from each scenario, and the keywords your AI can identify. You can't just drop the soft skills people expect by default, but an AI brings everything together, and you also expect it to be efficient, because every mistake can be costly. When you train an AI, it's like teaching it about human behavior and how to identify it through keywords and key questions. So to keep it premium: know your pain points, the key questions for your business, and how to help your AI identify the attributes that are non-negotiable.
Leo: So even in premium cases there's a mix of empathy and efficiency. How do you keep high accuracy and empathy while automating more?
Brandol Guerra: You can't lose accuracy for efficiency, that's terrible. The balance is to divide what your actual business pain point is versus what your people expect. Representatives are used to performing against KPIs, but an AI isn't conscious that it must perform, so you do that work for it: equip it by default to be accurate, connected to your articles and knowledge base. Accuracy is something the AI shouldn't have to worry about. It's a mix of your skills as the trainer and how well you know your business. If your material is accurate and has the right questions, keywords, and attributes, the AI spends less time identifying the problem or the frustration, and invests that time in the accuracy and efficiency you're after.
Leo: Rebag depends on trust, selling goods worth thousands. How do you decide what's safe to automate and what must stay human?
Brandol Guerra: AI can guide you through simple problems, troubleshooting, or how-to questions. But once it goes beyond your regular protocols and policies, people aren't truly open to speaking with a robot. So it's simple: if someone has invested too much time trying to resolve something, and the process has too much friction, that's the key, I like to reduce friction. AI has limitations today, it's based on its training and doesn't know how to handle someone who's frustrated or irate, and you're in jeopardy of losing that customer. So when there's too much friction, the AI should escalate to someone on my team, who can gather what the customer is looking for and continue where the AI left off. Escalations that might pull a customer to leave your business are when humans need to take place.
Leo: What workflows have you decided must stay human, and why?
Brandol Guerra: If you're trying to make sure a customer can purchase ASAP, that's a human. People are always in a rush, and AI can add friction because the protocol is strong. You don't want to automate someone who's willing to buy from your business. If they have a payment problem or don't understand how the platform works, you can't lose time, you step in. The other one is change management: when you launch something new, you can publish FAQs and train your people, but if customers don't understand what's next for them or how the business is evolving, a human needs to take over. Purchasing and change management are key, you don't fully delegate how your business grows to an AI.
Leo: You spent years at Sutherland and Teleperformance before moving in-house to Rebag. What's the one thing BPOs get wrong about customer service that you've been able to fix in-house?
Brandol Guerra: Those are great businesses, but in-house you can invest more time refining processes for accuracy and efficiency, and add value on the experience side. It's cliché to say we're not doing customer service but customer experience, yet it's real. Ten years ago people rushed to get rid of a call because your boss was watching handle time and NPS, and all of that sacrifices the essentials, especially attention to detail. It's like driving fast because you need to get somewhere, you risk a mistake. In-house you can create processes, streamline things, and train your team to understand why you're asking them to be empathetic and to follow a flow, because in a BPO every minute costs the business. If you don't explain the value behind every little thing, you might lose the trust of your people or a loyal customer over a misunderstanding.
Leo: You have a Six Sigma background, all about eliminating defects and improving processes. Can you give an example of a process you improved with Six Sigma that became a candidate for AI automation?
Brandol Guerra: Six Sigma is designed to eliminate waste, and waste means different things for different businesses. Say my top performer in productivity is doing tons of cases a day, but struggling with a process, they could be even more productive while keeping the best quality and CSAT or NPS scores. Six Sigma helped me find the gap. For example, an AI was gathering information from customers, but the agent didn't fully understand what the AI was getting, the problem was in the input. The AI captured that a customer was frustrated, but not the details of that frustration or what to do next, so the agent dug through the AI's output trying to decide the next step. Connecting eliminate-waste with how you connect information and reality, the fix was updating the AI prompts and what information it gathered, based on feedback from that specific agent who told me where they were wasting time.
Leo: You've led initiatives that boosted Google and Better Business Bureau ratings. In luxury, reputation is everything. What actually moved the needle?
Brandol Guerra: Understanding the right approach to read an interaction. In customer experience we tend to be reactive, asking what you thought after the interaction. But what moves the needle is noticing during the interaction whether you're already a detractor or promoter. My team is trained to read the attributes of an interaction within the first two or three minutes. If I can tell you're a detractor early, instead of taking you through the regular flow and putting you on hold to find an answer, I put you on hold to find the person who can help. And second, timing: if we couldn't resolve it in time, our team alerts us after the interaction and we call the customer back. About 50% of the time people aren't upset because something couldn't be resolved, they're upset because they couldn't express what the problem was. Once they cool down, an agent helps them shape what they need. There's a big difference between what a customer wants and what they need, and identifying it early triggers a good resolution. It's not a sin to say let me circle back, but many people let customers go because they're hesitant to ask for time or worried about productivity.
Leo: There's an Einstein line that with an hour to solve a problem he'd spend most of it identifying the problem first. You've led teams through big transformations, and people often see AI as a threat. How do you get your team on board without fear?
Brandol Guerra: It's about reducing resistance to change. First, train them to take advantage of AI, and it comes back to what you want versus what you need. If I want my chats to finish in two minutes, the real question is what I need to do that. The same applies to you as a leader: if you want to increase performance 20% next quarter, the question is what you need to do, and that's where AI takes place, trained on your company's reality, front-facing and internally. Many people use ChatGPT for basic questions, but it doesn't know your background. My team is equipped to the business reality, they know what's happening, and when we make a change I make sure they know in time so they understand the need behind it. It works backward: train your team to understand what they need before what they want. And involve them, because successful teams aren't a one-person army. Your whole team can help train the AI, and then they know what it's trained for and how to use it properly.
Leo: Imagine a head of customer service at a luxury brand has been told to implement AI. Where should they start, and what should they avoid?
Brandol Guerra: Start by foreseeing the challenges of change, because this industry is very dynamic, with brands constantly launching new references and specifications. Authentication is a separate matter, but if you want to implement AI you need someone or a team behind it who stays up to date, masters of research, because AI can do the trainer's work but you have to catch the details, you might lose valid references or carry ones you no longer need. The key is timing: if you and your team are up to date, your AI won't fall behind.
Leo: Rapid fire. Speed or accuracy, which matters more?
Brandol Guerra: Accuracy. You'd never sacrifice quality for quantity.
Leo: True or false, within three years most luxury brands will use AI for first-line customer interactions.
Brandol Guerra: True. It opens room to invest more time on the important matters. AI is the first line of defense, then your team is the core.
Leo: Finish the sentence: the company that wins in 2026 isn't the one with the best AI, it's the one with the best...
Brandol Guerra: Adaptability. We're humans in a universe that's always changing, and if you don't resist change you can adapt to anything. That's the basics of evolution.
Leo: Brandol, this was great. Where's the best place for people to follow your work?
Brandol Guerra: On LinkedIn, that's where I'm most connected with my network and community. You never know who's your next mentor, friend, or contact for a great deal.
Leo: Perfect, Brandol. Thanks for joining us.
Brandol Guerra: Thank you for your time. See you next.
Leo: And for everyone listening, if you want more honest conversations about what it takes to implement AI in customer service, make sure you subscribe to the Fini Podcast and we'll see you next time.
How do you deploy AI in a luxury customer service environment?
Brandol Guerra starts by defining the attributes that make service premium, then trains the AI on the key questions and keywords that detect them, like recognizing when a customer is upset. He grounds the AI in a strong knowledge base so it is accurate and efficient, treating it as an extension of the team rather than a generic chatbot.
What should stay human in luxury support?
Brandol keeps humans on anything where friction risks losing the customer: a shopper ready to purchase, a broken payment or checkout, and change management like explaining how the business is evolving. AI handles troubleshooting and how-to questions, then escalates when a customer has invested too much time.
How does Six Sigma apply to AI in customer service?
Six Sigma targets waste. In Brandol's example, an AI flagged that a customer was frustrated but not the details or the next step, so an agent wasted time interpreting it. The fix was rewriting what the AI was prompted to capture, based on feedback from the agent using it.
What's the difference between BPO and in-house customer experience?
In a BPO every minute has a cost, so teams rush to end interactions, sacrificing attention to detail. In-house, you can build processes and teach the team the value behind empathy, attention, and the interaction flow, shifting from customer service to customer experience.
