Latest / The CMO Podcast with Fexingo: Marketing Leadership, Budgets, and Executive Strategy / How CMOs Are Using Branded AI Assistants for Customer Service
Transcript
- Lucas: If these marketing conversations have sparked something you've actually used — maybe a new way you're thinking about brand utility or customer touchpoints — we hope this next one lands with that same practical edge. Luna: Alright, what's the hook? Lucas: So it's mid-2026. The chatbot era is basically over. Customers have learned to spot a scripted bot from a mile away. But what's replacing it, especially for CMOs, is something much more interesting: branded AI assistants that aren't just answering questions but actually embodying the brand's voice and personality. Luna: You're talking about AI that doesn't just respond but represents — like an extension of the brand team. Lucas: Exactly. And the example I keep coming back to is Domino's Pizza Enterprises. They just launched a generative AI ordering assistant named 'Dom' across Australia and New Zealand. This isn't a menu bot that reads back your order. It's a conversational agent that can handle complex requests — 'I want the same as last Friday but with extra pepperoni and swap the drink for a Sprite' — and it does it in a tone that's consistent with Domino's cheeky brand voice. Luna: Right. And the numbers are pretty striking. I saw that Dom is handling over 200,000 orders per month, and customer satisfaction scores are hovering around 95 percent. Lucas: That 95 percent is the number that jumps out. Because if you think about it, a customer service interaction that scores that high is rare even with human agents. So Domino's has essentially created a brand touchpoint that's more reliable, more consistent, and arguably more on-brand than a rotating cast of call-center employees. Luna: But is this really a marketing play, or is it just a smarter ordering system? To me, that sounds like operations. Lucas: I'd argue it's both, but the marketing angle is the one that CMOs need to pay attention to. Because every time a customer interacts with Dom, they're not just placing an order — they're reinforcing the brand relationship. Domino's has always positioned itself as fast, friendly, and a little bit irreverent. Dom's language mirrors that. It uses slang, it cracks jokes, it remembers your past orders. That's brand building through utility. Luna: So it's not just about reducing friction in the order flow — it's about making the friction reduction itself feel like a brand moment. Lucas: Exactly. And it's also a massive data play. Every conversation generates zero-party data — things the customer explicitly tells you about preferences, frequency, even life events. 'I'm ordering for a party' or 'I'm feeding the kids after soccer.' That kind of data is gold for a CMO because it lets you personalize marketing without relying on third-party cookies. Luna: But there's a risk, right? The same consistency that makes Dom great could backfire if the AI says something off-brand or, worse, offensive. Lucas: That's the big fear. Domino's spent months training Dom on their brand guidelines and had human reviewers in the loop for the first few weeks. But the thing is, generative AI is probabilistic. It can still hallucinate or respond in unexpected ways. So the CMO has to decide: how much autonomy do we give the assistant, and where do we draw the line? Luna: I know some brands have solved that by having the assistant escalate to a human the moment it senses a negative emotion or a complex complaint. Lucas: That's a smart guardrail. And it's actually what Dom does — if the conversation goes off-script or the customer seems frustrated, it hands off to a human agent seamlessly. The key is that the handoff doesn't feel like a failure; it feels like the brand cares enough to get you the right help. Luna: So what's the takeaway for a CMO in, say, retail or financial services who's considering a branded AI assistant? Where do they start? Lucas: First, they need to define the assistant's personality as meticulously as they define their brand's visual identity. That means writing a tone of voice document that covers not just what the assistant says but how it handles edge cases — like a customer swearing, or a customer asking for something the brand doesn't offer. Luna: And that's harder than it sounds because you're not scripting every line. You're training a model on examples, and the model will generalize. Lucas: Right. So you need a really robust set of examples that cover the kinds of scenarios your customers actually encounter. And then you need a testing phase that's longer than you think — maybe three to six months — before you put it in front of real customers. Luna: I also wonder about the cost. Building a custom AI assistant with generative capabilities isn't cheap. You need the LLM, the integration with your backend systems, the training, the guardrails, the monitoring. Lucas: True. But if you look at the ROI, it can be compelling. Domino's didn't just save on call center costs. They saw an increase in average order value — because Dom is good at upselling. 'Would you like to add a dessert for two dollars more?' That kind of thing. So the assistant becomes a revenue driver, not just a cost center. Luna: And that's the kind of argument that gets a CMO a seat at the table when the budget is being discussed. Lucas: Exactly. One thing we haven't mentioned is the role of voice. Dom is text-based right now, but plenty of brands are experimenting with voice assistants that sound like the brand. Think about a hotel chain whose voice assistant sounds warm and professional, or a car company whose assistant sounds sophisticated and knowledgeable. Luna: Voice adds another layer of complexity because now you're dealing with tone, accent, pacing. But it also deepens the brand connection. Lucas: And that's where I think we'll see a lot of innovation in the next 12 months. We're already seeing CMOs in hospitality and retail piloting voice assistants for concierge services and in-store assistance. The technology is getting good enough that the uncanny valley is shrinking. Luna: Before we wrap, I want to circle back to something you said earlier about brand relationships. Do you think customers actually form a relationship with an AI assistant? Or is it just a tool they tolerate? Lucas: I think it depends on the execution. If the assistant is purely transactional — 'Your order will be ready in 15 minutes' — then no, it's a tool. But if it remembers that you always order a large pepperoni on Friday nights and says 'Hey, same as usual?' with a friendly tone, that creates a sense of recognition. That's the beginning of a relationship. Luna: And that recognition is exactly what brands are trying to create at scale. Lucas: Right. And the data shows it works. Domino's reported that repeat order rates are higher among customers who use Dom compared to those who order through the website or app without it. So the assistant is actually driving loyalty. Luna: That's a powerful metric. And it's one that CMOs can use to justify the investment. Lucas: Speaking of investment — we talk a lot on this show about the tools and strategies that CMOs are betting on, and we try to keep the conversation as practical as possible. We deliberately don't run ads on these episodes because we think the content should stand on its own. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. Luna: Yeah, it's a small way to keep the podcast ad-free and focused on what matters. Lucas: So back to Domino's — I think the bigger lesson is that AI assistants are becoming a new brand channel. They're not just customer service; they're a place where the brand lives and breathes every single day. Luna: And that means CMOs need to be deeply involved in designing them — not just signing off on the tech. Lucas: Exactly. The CMO should be the one asking: 'What does our brand sound like? What does it remember? What does it apologize for?' Those are marketing decisions, not IT decisions. Luna: So the question for our listeners is: if your brand had an AI assistant today, would customers recognize it as yours? Lucas: That's the right question to walk away with. For The CMO Podcast with Fexingo, I'm Lucas. Luna: And I'm Luna. See you next time.