Every leadership team I talk to is having the same conversation about AI, usually in the same tone of voice. The pressure to adopt is real. The flood of tools is real. And underneath both sits the quiet worry nobody puts on a slide: move too fast and you cheapen what you've built, move too slow and you get left behind. Most premium brands resolve this tension by freezing. The question was never whether to use AI. It's where, how, and under whose judgment. The answer I give them is a framework I call the House Line.
The playbook you've already been pitched
There's a standard answer, and you've probably heard it. The AI companies and the big consultancies converge on the same six moves: get a sponsor at the top, pick high-ROI use cases, upskill the team, wrap it in governance, measure outcomes, iterate and scale. None of that is wrong. It's just built for a different kind of company, one whose advantage is efficiency.
A brand-led business has a different problem, because its advantage is meaning. McKinsey and QuantumBlack put it bluntly in 2025: AI flattens positioning, eroding the discretion, scarcity, and high-touch service that underpin luxury value. Read that again, because it names the actual risk. The danger for a premium brand is not that AI makes the work faster. It's that AI makes the brand less authored.
A use-case list can't protect you from that, because use-case lists don't know which of your workflows carry the brand and which merely keep the lights on. That distinction is the whole game.
Two kinds of work
Inside any brand-led company, every workflow is one of two things: brand-defining or brand-enabling. Brand-enabling work keeps the machine running: inventory reports, compliance, routine trade email, the endless choreography of status updates. Brand-defining work is the reason anyone pays a premium. It carries one of five signatures: taste, the judgment that picks this and rejects that. Authorship, a discernible hand behind the work. Symbolic meaning, what the choice stands for beyond what it does. Narrative continuity, today's work extending a story that started long before it. And emotional distinctiveness, the feeling only your brand produces.
The test is one sentence: if the work could belong to any brand, it doesn't carry the brand. A sales dashboard could belong to anyone. Hand it to AI tomorrow. The blend, the voice, the welcome a guest gets at the estate, that work is the brand, and it stays human.
The House Line
That split is one axis of a simpler map, and the map is the framework. The House Line sits on top of positioning, not instead of it: once a brand's position is locked, it decides where AI belongs. Draw two axes. The horizontal one runs from brand-enabling work to brand-defining work, the distinction we just drew. The vertical one runs from algorithmic reliability at the top, work that is repetitive, verifiable, and has one right answer like a math problem, down to human judgment at the bottom, work that runs on taste and discernment and can't be checked the same way. Where a piece of work lands on that map tells you what to do with it.
Brand-enabling, reliable. Research, sales analysis, filings. Maximize AI and scale it.
Brand-defining, reliable. AI works inside a taste constraint. A human signs off.
Brand-enabling, judgment. A sensitive negotiation, a hiring call. AI assists, a human decides.
Brand-defining, judgment. Human craft, the reason they pay the premium. It stays human.
Three of those corners are straightforward. Maximize AI on the reliable, brand-enabling work. Augment the reliable, brand-defining work with AI held inside a taste constraint, a brief tight enough that the model stays on-voice, with a human making the final call. Let AI assist on the judgment work that doesn't define the brand. The corner that matters most is the bottom right, where brand-defining work meets human judgment. That is the sovereign space. It is the craft a customer is actually paying for, and it does not get optimized. It gets protected, loudly.
The four tests
Placing real work on that map by instinct is how companies end up with scattered pilots and no line anyone can state. To find the human-judgment edge, I run every workflow through four questions.
Identity. If AI did this work alone for a year, would the brand still feel like itself? If the honest answer is no, it stays human.
Taste. Does subtle judgment matter more here than speed or correctness? Tasting notes can survive a competent first draft. Choosing which of forty drafts sounds like the house cannot. Taste-sensitive work stays human.
Reversibility. If it goes wrong, can it be fixed without brand damage? A mislabeled spreadsheet is an apology. A generic brand voice in market for a quarter is an erosion. Irreversible work stays human.
Scarcity. Does limiting the output increase its value? Some things are valuable precisely because there aren't many of them. Anywhere scarcity is the point, a human stays at the gate.
Work that passes all four gets delegated without guilt. Work that fails any of them stays on the human side of the line, with AI permitted to assist but never to own. The principle underneath: advance authorship, industrialize execution.
A territory, not a blank page: the taste constraint
Delegation is half the discipline. The other half is the taste constraint, the standard AI-touched work has to meet, and most companies never write it down. AI shouldn't get a blank page. It should get a bounded territory: the house's tone rules, visual lexicon, narrative canon, cultural red lines, forbidden generic language, and the thresholds where a human takes over. Not quality control. Taste, written down.
You already know what work that fails this standard looks like, because the internet is currently drowning in it. Technically fine. Instantly generic. It could describe anyone's product, anyone's estate, anyone's story. The bar is simpler and much higher: specific, true, unmistakably yours. Good enough to carry your name, or not at all. I've written before about why positioning is the real AI strategy. This is that argument made operational.
Your best people aren't resisting. They're protecting something.
Here's the part the rollout decks get wrong. When skilled people go quiet in the AI workshop and then change nothing, it gets filed as a training problem. It almost never is. Bain and the Comité Colbert found in 2025 that the great houses stay deliberately cautious about AI in creative functions, unwilling to risk altering luxury's essence or compromising authenticity. That's not technophobia. That's professional pride doing exactly its job.
Your best people resist AI when it appears to cheapen the judgment they're proud of. They don't need convincing that the technology is powerful. They need to believe it will protect and expand the work only they can do. And the honest math is on their side: Asana's research across ten thousand knowledge workers found that only about 27% of the day goes to the skilled work people were actually hired for. Most of the rest is work about work, chasing documents, status updates, switching between apps.
That's the trade worth naming out loud. AI absorbs the work about work, inside the taste boundaries, below the line. The day goes back to judgment, authorship, care, and craft. The dividend isn't time saved. It's judgment returned. Same hours, more craft.
Every great house knows its House Line
Luxury was never the absence of technology. It's the presence of judgment. The houses that get this right won't be the ones that adopted fastest or held out longest. They'll be the ones with a drawn line the CEO can repeat to the board: what AI absorbs, what it assists, what stays human, and the standard everything AI-touched has to clear before it carries the name.
If your company can't state its House Line in a sentence yet, that's not a tooling gap. It's a brand decision waiting to be made. Start by drawing it. The House Line is the AI-delegation framework inside my broader brand positioning and strategy methodology, and it only holds when the positioning underneath it is sharp.
