Omotenashi: The Japanese Philosophy Behind Better CX and AI

Recently I had the chance to talk with Kane Simms from VUX World, and we talked about omotenashi, CX, and conversational AI. It was a great conversation on how a centuries-old Japanese philosophy of hospitality can shape the way we design AI-powered customer experiences.

In this article, I’ll walk through the main areas we touched on during our conversation.

I’m originally from Spain but have called Japan home for 12 years, after calling Edinburgh, Dublin, and Shanghai home. Learning Chinese in Shanghai is actually what led to a job offer in Tokyo.

My path into conversational AI ran through business development, then retail tech, then five years at Rakuten building customer service chatbots across its many verticals, and then going through the shift from pre-LLM to generative AI.

Omotenashi is a Japanese philosophy of putting the customer first: anticipating their needs, apologizing sincerely when something goes wrong, and solving problems at the right moment. Actually, it’s not different in kind from other good service elsewhere in the world, but I would say the Japanese can bring it to another level.

Small examples stick with me: supermarket staff bowing to the warehouse every time they enter or leave, even when no one’s watching; Shinkansen cleaners bowing to the train car before moving to the next one. It’s become a market differentiator too, Japanese airlines and car companies actively market omotenashi as part of their brand.

Japanese customer service places huge weight on apologizing sincerely before even attempting to fix the problem. Studies show customers often value a sincere apology more than compensation, the opposite of what I’ve generally seen in the West, where the fix usually comes first.

This extends to how mistakes are handled generally: deflecting blame, even partially, is a real misstep. The expected response is closer to “this is on me, I’m sorry”, putting both parties on the same footing before moving to a resolution. It may connect to how much more individualist Western culture is by comparison; a hedge like “sorry you feel that way” just doesn’t land the same way in Japan.

Japan consistently scores among the lowest average NPS across industries compared to markets like Brazil, India, or Australia, not because service is worse, but because expectations are higher. The same restaurant often scores lower on a Japanese review platform than on Google Maps.

That gap shows up directly in conversational AI design. Japanese has 10-15 ways to say “sorry,” and choosing the right level of formality for a situation matters; generative AI now makes that calibration possible. Escalation matters just as much: Japanese customers strongly prefer a human when a bot can’t help, and 56% say they won’t return to a brand after one bad interaction.

Anticipation is core to omotenashi, sensing a need before it’s voiced and proactively reaching out, rather than waiting to be asked. That’s as much a design principle as a feature: mapping likely problems by contact type and building in proactive signals wherever possible.

Language adds complexity. Japanese often avoids direct refusal in favor of ambiguity (aimai), used to preserve harmony; words like chotto (”a little”) or demo (”but”) can signal dissatisfaction far more subtly than an outright complaint. Catching that requires prompting AI with specific examples of these soft expressions. Even in person, a phrase like “let me consider it” (kentoshimasu) is, in the right context, a clear no, it just takes time living in the culture to read reliably.

Expectations are simply higher here, so metrics that look weak by other markets’ standards may just reflect a tougher baseline. Beyond apology tone, linguistic etiquette matters: Japanese uses three writing systems, and mixing them inconsistently within a conversation is something native speakers notice immediately. Keeping messages under ~50 characters helps too, since the script is visually denser to read.

Digital personification is a distinctly Japanese opportunity, mascots are everywhere, and giving a chatbot its own named, personified identity tends to land well. Integration with local tools matters too: LINE, Japan’s dominant messaging app, offers deeper business functionality than WhatsApp does in the West, making it close to a must-have channel.

This isn’t about Japan having answers the West hasn’t found, there’s plenty Japan could learn from a more direct approach too. The interesting question is what each side can take from the other.

What omotenashi offers is a genuine raising of the bar: more accountability, more sincerity, and a habit of pausing to empathize before jumping to the fix. Solving the problem is a given, the harder, more valuable part is pausing first to say “I understand this is frustrating, and now let’s fix it.” That’s the piece worth building into how we design conversational AI everywhere.

Hello and thanks for reading. If this is your first time here, I’m Daniel, and I’ve spent 12 years living and working in Japan, leading Customer Experience and Conversational AI for some of Japan’s biggest companies. I write about what gets lost when technology moves fast, and what Japanese service culture can teach us about bringing it back.

I publish an article weekly, next one coming soon. In the meantime, you might find the following reads interesting: