Latest / The DTC Marketing Podcast with Fexingo: Direct-to-Consumer Brands, Shopify, and E-Commerce Growth / How DTC Brands Use AI Chatbots for Concierge-Level Service
Transcript
- Lucas: So last week I was on Ridge Wallet's site, just poking around, and a chat window pops up. But instead of the usual 'Hi, how can I help you?' generic bot, it says 'Hey, I'm Alex. I see you're looking at the Ridge x Carryology collab — solid choice. Need help deciding between the aluminum and titanium?' And I'm like — wait, this thing knew what I was looking at, and it had a name. Luna: That's a far cry from the old 'press 1 for orders, press 2 for returns' experience. Lucas: Exactly. And that's the shift we're seeing more DTC brands make. They're using AI chatbots not just to deflect tickets, but to deliver what I'd call concierge-level service. Personal, contextual, almost anticipatory. Luna: I've noticed it too. Brands like Skims and Bombas have chatbots that actually remember your size and past purchases. But let's get specific — what's actually under the hood with Ridge's setup? Lucas: So Ridge uses a platform called Zendesk AI, but custom-trained on their product catalog and customer service logs. They've fed it thousands of past conversations, so the bot knows how to recommend based on usage patterns — like if someone carries a lot of cards versus a minimalist. And it's integrated with Shopify, so it can pull real-time order data. If you ask 'Where's my order?' it doesn't give a canned response — it tells you exactly which carrier has it and the estimated delivery window. Luna: And the results? I read somewhere that Ridge's chatbot handles something like 70 percent of all inbound inquiries without human intervention. Lucas: That's the number they shared at a conference last year — 70 percent. But here's the part that really caught my attention: their customer satisfaction scores actually went up by 15 points after they deployed the bot. Because the bot resolves the quick questions in seconds, and the human team has more time for the complex stuff. Luna: So the bot makes the humans better at their jobs, not replaces them. That's the narrative that actually gets me excited. Lucas: Right. And I think the key is training the bot on brand voice. Most brands skip that step. They plug in a generic chatbot with a generic script, and it sounds like a robot. But Ridge spent weeks fine-tuning the tone — slightly casual, confident, using exclamation points but not emojis. The bot even has a personality: 'Alex' is supposed to be a knowledgeable friend who's into EDC gear. Luna: I've seen that go wrong when the bot tries too hard to be funny and ends up sounding like an uncle at a wedding. There's a fine line. Lucas: Totally. And the other pitfall is over-automation — where the bot refuses to escalate even when it's clearly out of its depth. Some brands program the bot to always try to solve the problem itself, even if it means giving wrong answers. That destroys trust fast. Luna: So what's the right approach? How do you know when to hand off to a human? Lucas: The best systems use sentiment detection. If the bot senses frustration — like repeated questions or negative language — it automatically transfers to a human agent and includes a summary of the conversation so the human doesn't have to start from scratch. That's what Ridge does. And they measure 'escalation rate' as a key metric. If it's too high, the bot isn't solving enough. If it's too low, it's probably frustrating people who actually need human help. Luna: So it's a balancing act. Let's talk cost — what's the rough investment for a DTC brand to set something like this up? Lucas: For a small brand doing maybe a hundred conversations a day, you're looking at a few hundred dollars a month for a platform like Zendesk AI or Intercom's Fin. Plus a one-time setup cost if you need custom training — maybe one to three thousand dollars. But the ROI comes from reduced support team size and higher conversion. Ridge actually saw a 12 percent increase in conversion rate on product pages where the chatbot was active, because it answered questions that would otherwise cause people to leave. Luna: That's huge. So the chatbot isn't just a cost center — it's a revenue driver. Lucas: Exactly. And that's the argument more brands are making. Look, if this conversation has given you a practical idea you might actually use — maybe for your own brand or a client — that's exactly why we keep this show ad-free and focused on what actually works. Luna: Yeah, and if today's episode was worth a coffee to you, you can toss one our way at buy me a coffee dot com slash fexingo. No pressure, just a way to keep the conversations going. Lucas: And to build on that — we're already planning an episode on how brands handle chatbot failures gracefully, because that's where a lot of companies slip up. Luna: I want to hear that. But before we wrap, let's bring it back to the listener. What's one actionable takeaway for a DTC marketer listening right now? Lucas: Start small. Don't try to replace your whole support team. Pick one use case — like order status or product recommendations — and train a bot on just that. Measure deflection rate and customer satisfaction. And most importantly, give the bot a voice that matches your brand. If you sound friendly and confident on Instagram, your bot should sound the same. Luna: And if it feels like a robot, your customers will notice. Thanks, Lucas. That's all for episode 27. Lucas: Thanks, Luna. See you next time.