Strategy
What the agentic advantage actually is in this business, where the moat sits, and what the company will not pursue.
What you are buying is access to one person across seven areas, from the strategy down to the code itself, standing on twenty-two years of Nordic retail. Below: those seven areas, five ways to buy them, the method behind all five, and five habits that differ from how a firm would do the same job.
In a firm these sit in different practices and you are passed between them. Twenty-two years in Nordic retail is not one of the seven, it is the ground they all stand on.
What the agentic advantage actually is in this business, where the moat sits, and what the company will not pursue.
The organisation designed around agents rather than an existing org chart automated. Decision rights and accountability.
A stable core and a tool layer meant to be replaced. Where the build-or-buy line falls, and why.
Agents where work repeats, code where rules are known, people where judgement lives. Ownership and quarterly review.
Leading people who have never worked this way, and building the capability bridge deliberately instead of waiting for it.
Fifteen years of marketing and marketing automation, now the most agent-native discipline there is.
Businesses that did not exist before, taken from nothing to trading, including the parts AI is no use for.
Twenty-two years. Assortment, buying, campaigns, shrinkage, staffing, stores and e-commerce in the same business.
The same access, bought in five sizes. Each one names what lands on your desk, how long it takes and what it costs, before you have to ask.
A written read of where you actually are, delivered in the same week.
Not a maturity score, and not a proposal for a larger engagement.
One agent in real production, with the person who will keep it running already trained on it.
Not a proof of concept, and not a demo that needs me present to work.
Someone in the leadership team while the change is actually made, ending with a team that runs it without me.
Not a steering-group seat, and not advice nobody has to act on.
A field report from inside a company that did it, including what it cost.
Not a trend deck, and not a talk that ends in a sales pitch.
Hours, on whatever the problem actually is, with no shape imposed on it.
Not a retainer. Hours are booked and used, never reserved.
The moves are deliberately dull. What makes an AI programme stall is almost never the model, it is that nobody decided who is accountable when the machine is wrong.
Walk the work from the inside, task by task, not from a system diagram. What runs, what is a pilot, and what is theatre.
Agents on one side, ordinary code in the middle, people where context and judgement live. In writing, with a name against each one.
One thing into production with its guardrails, its logging and its monitoring. Narrow on purpose, because a narrow thing finishes.
A named owner, trained, with the code in your repository and a standing quarterly decision to keep the agent or switch it off.
The large firms are good at this work. These are the five places where a one-person practice can simply behave differently, and does.
| What a consultancy does | What Young Digital does |
|---|---|
| A team you meet at the pitch, and juniors on the work | One person. The one in the room is the one doing it, and there is nobody to bill underneath |
| A price arrived at after they have judged what you can pay | €250 an hour, published, the same number for everyone, on the public website |
| The deck is the deliverable, and the pilot ends when the invoice does | It runs when I leave, it is in your repository, and your named owner has already kept it running once |
| Case studies with the failures edited out | Seven principles published free, each with the case where it breaks, and two that have no counter-case at all |
| A retainer, renewed | Refused. Advice nobody has to act on is a subscription, not work |
Two hours costs €500, and it comes off the first invoice if it turns into work. If it does not, you have had two hours from someone who has already run the thing you are about to attempt.