AI Diagnostic
For teams who know they should be doing something and want to know exactly what.
- Stakeholder sessions across the business
- Opportunity map, ranked
- A prioritized roadmap you can act on
You've talked about AI for months. I help leadership teams pick the one thing worth doing, build it in six weeks, and make sure people keep using it after I leave.
ScrollEveryone's using ChatGPT. Nobody's using it the same way.
Twenty years of beautiful work, and nobody can find any of it.
You bought the licenses. The licenses are fine. The habits haven't changed.
Often not a technology problem but a "who has time to figure this out between client calls" problem. That's the bit I fix.
I start wherever you are. If you already know where it hurts, we go straight to a diagnosis. If you don't, we do a proper discovery first. Either way, here's what happens.
Short conversations with the people who do the work, or a fuller discovery if you're earlier on. A look at the tools and files you already have.
We pick one high-value use case and tie it to something you care about: hours back, revenue, or how clients feel.
A focused sprint, usually six weeks, with a small pilot group using the thing on real work from week two.
Training, a couple of in-house champions, short playbooks and simple measurement.
Senior strategy plus a small hired build team when we need one. You get the thinking and the doing, without paying for a floor of consultants.
For teams who know they should be doing something and want to know exactly what.
One priority use case, taken from scoping to a working pilot, with build support included.
A hand on the wheel through rollout, for leadership teams who want someone in the room.
Owners and leadership teams of 20 to 300 people, where the work is human, beautiful and detail-obsessed.
A searchable archive of every past design, by look, venue and season.
Itinerary and proposal drafts in your voice, ready for a human finish.
Guest notes, briefs and handovers that don't live in one person's head.
Sell-in decks and product copy drafted from what you already know.
Faster first rounds, so the team spends time creating and thinking.
I spent fifteen years in luxury brand, experiential and retail marketing, mostly agency-side, looking after big portfolios and the executives who sponsored them.
Then I joined an AI startup as a founding operator. I started on growth, but kept getting pulled into what happened after the sale, because that was where projects stalled. So I moved into transformation, and ended up running it for clients in travel, wine and events. The problem I see everywhere now is that companies have bought the AI tools and their teams still aren't using them. Most of my job is helping busy people change how they work on an ordinary Wednesday afternoon, when the old way is still right there and much more comfortable.
I'm South African born and these days split my time between Brooklyn, London and Cape Town. Outside of work I'm happiest choosing and arranging things. I've worked as a freelance stylist, and I curate everything else for free whether anyone asked or not: a dinner table for ten, a trip planned down to the last stop, the one object that makes a room.
Recognition Technical partner on an award-winning AI project.
Dani
Most AI help stops at strategy or at the software. Adoption is the part after that: getting your people to use the tools in their actual work, and keep using them. I pick one useful thing with you, build it with your team, and make sure it sticks.
Owners and leadership teams of 20 to 300 people, mostly in luxury travel, events, hospitality, premium brands and creative agencies. If your business runs on taste and relationships, and nobody has time to figure out AI on the side, it's probably you.
Training shows people what the buttons do. I start from your real work: one workflow, built with the people who do it, used on live jobs from week two. Training is part of it, just not the whole of it.
The Diagnostic takes 2 to 3 weeks. The Sprint takes 6 weeks, with a pilot group using the thing from week two. Partner in Residence runs month to month, for as long as it's useful.
Usually, yes. And that's completely normal. Nobody I've worked with had tidy files. We scope around it: start where the data is good enough, fix what the pilot needs, and say plainly what a cleanup would take.
Maybe not. Often the tools are fine and the problem is that nobody's shown the team what good looks like for their actual work. I'll tell you if something new would earn its keep.
NDAs as standard. No client data used to train models. Access scoped to what the work needs, and removed when we're done.
Both. I lead the strategy and adoption, and bring in a small build team for the sprint. You get working software, not a recommendations deck.
Each engagement is fixed-fee and scoped up front. We'll talk numbers on the first call so nobody wastes time.
Brooklyn, London and Cape Town, depending on the month. I work remotely or onsite, and I'm happy to get on a plane for a workshop.
Thirty minutes, no slides. Tell me what you've been meaning to do with AI and I'll tell you honestly whether I can help.
Thanks, there.
Got it. I'll reply within 24 hours, from whichever time zone I'm in. In the meantime, you can skip the wait and book a 30-minute call.
Book a 30-minute call