In February 2026, I ran an AI-driven guest qualification system targeting HNWI travellers at Nara. The brief was built from two years of CRM data: spend profile, recency, frequency of engagement, lifetime value. AI extracted the patterns. I built the brief and set the channel strategy. The creative team executed against it.
The campaign did not feel like AI content. It felt like a team with a precise picture of who it was talking to. February was our strongest month of the season: footfall up 33 per cent on January, bookings nearly doubling month-on-month, and across the group, the best Valentine’s season in the venue’s five-year history, up 50 per cent year-on-year.
The diagnosis circulating in marketing circles is that AI has lowered the quality of thought leadership. That is the wrong diagnosis. What has lowered quality is not the tool. It is the substitution: AI replacing the critical thinking rather than sharpening it. The difference is a mental model.
Think of it like a professional kitchen. AI is exceptional prep, capable of mapping the mise en place and identifying which flavours have already been overused in the market. But no restaurant has ever served guests a dish the prep list designed. The chef still cooks, and the judgment going into the plate is earned through years of service and experience, not built through prompts.
Generic content happens when AI is treated as a faster way to write. Better content happens when it is treated as a faster way to think: to research, synthesise, and frame the argument before a single line is written.
I use AI for the inputs: audience pattern analysis, research synthesis, and data work that would otherwise take my team a week. For our February campaign, AI processed two years of booking data and surfaced a pattern we had not formally named: 30 per cent of our guests were booking within three days of their visit. No regular CRM cycle catches intent at that velocity. The brief came from that pattern, and so did the audience profiling and the qualification criteria. None of this is AI authorship. It is AI as infrastructure.
Hospitality makes this distinction unusually clear. Most industries allow thought leadership to rest on principle alone. A framework developed through analysis carries weight. A strategic observation derived from data holds up. In hospitality, neither is enough on its own. Every meaningful insight traces back to a specific property, a specific guest’s emotion, a decision made under real operational pressure.
A marketing director at a Dubai F&B group cannot write credibly about AI in guest experience without grounding the argument in something that actually happened: a campaign brief torn up and rebuilt in week two of a soft launch, a pricing call made when demand fell overnight, nine people executing against a deadline that was not going to move. That is the knowledge base the argument draws from. AI does not have access to it. It cannot be transferred through a prompt.
As AI tools become more capable, the shortcut becomes more tempting, and the outputs sound more polished, more confident. They are also progressively more identical. Built from the same training data and shaped by the same feedback loops, they converge on outputs that sound authoritative and read as interchangeable. Marketing ends up with strategy and tactics that sound right and land nowhere.
AI is, by design, an agreeable collaborator. It does not challenge a premise. It develops one. Feed it a half-formed argument and it returns the same argument, more articulate, better structured, harder to doubt. A practitioner who spends enough time in that loop starts to mistake a refined idea for a tested one. The thinking feels sharper because the prose reads better. It is not sharper. It has simply been reflected back at a higher resolution. That feedback loop is addictive, and it is why so much AI-assisted thought leadership sounds confident while saying nothing that would survive a serious challenge in the room.
Hemingway called it the iceberg theory: the dignity of the iceberg comes from the seven-eighths beneath the surface. AI generates the visible portion efficiently, but the marketer’s years in the room are the submerged mass, and readers feel its absence when it is not there.
Use AI to process what you have. Use it to find the pattern you have not named yet. Then write from the part of that pattern that only you understand.
The practitioners who will stand out are not the ones who avoided AI. They are the ones who used it to prepare better, and then wrote the part only they could write.
By Lucas Stamm, Director of Marketing, Digital & Innovation, Nara Hospitality Group.
