Your role and your tasks are not the same thing.
Agents take tasks. People keep the role: the judgement, the relationships, the accountability. Separate the two and most of the fear about AI goes with it.

We help organisations understand what’s changing, decide what deserves their attention, and build the agentic operations around it. With their people, not instead of them.
This is how a transformer reads a sentence. Each word weighs every other word, then rebuilds its meaning from the ones it attends to. Tap any word to see how the model reads it. Organisations need the same discipline.
In June 2017 a paper called Attention Is All You Need changed what machines can do. Two minutes on what it means for how organisations work.
Leaders can’t direct what they don’t understand, so we start in the room, with football’s governing bodies, women’s sport, business schools and leadership teams. Then we stay and build: the frameworks, the agent systems and the new ways of working, until your people run them.
AFC Conference · Tokyo
RAISE · Emirates Stadium
Ask an AI where to spend a £250,000 marketing budget. Women’s sport wasn’t in the answer. Your next customer may be a machine.
RAISE · Women’s sport
Forty seven associations, every size and starting point. One plan each could begin the following week.
AFC · Tokyo
Women’s sport doesn’t need to catch up. It can skip the queue.
RAISE · In conversation
The price of intelligence is collapsing. What you point it at is the decision.
AFC · Tokyo
One real question, asked live, and the answer read out loud.
RAISE · The test
Give your people a pitch, not a cage. Clear lines on what never goes to an agent, freedom inside them.
AFC · TokyoEvery organisation produces more than it could ever attend to. Most of it doesn’t need a person at all.
Scroll, and watch what deserves attentionAgents can carry the volume. People should carry what matters. The work is deciding which is which, and then building it properly.
In June 2017, eight researchers at Google published the paper behind every modern AI system. Its insight was simple: a model works best when it decides what to attend to, and ignores the rest. Organisations work the same way. So we named the studio after the paper.
arXiv:1706.03762 · Vaswani et al. · 17 is the year, 06 the monthYou’ve seen a model weigh a sentence. This is the same mechanism weighing an organisation: one question, every activity, and a fixed amount of attention to share between them. Most get almost none. Choose a question.
Illustrative weightsThe models are becoming a utility. When every firm can rent the same mind, what’s left to compete on is what you choose to point it at.
From Ground Truth · Everyone Is About to Have the Same Intelligence ↗Most AI so far has sped up the old operating model. The same meetings, the same handoffs, the same reports. Only quicker.
From Ground Truth · We Made Work Faster ↗Counting how many people use the tools misses the point. The bigger prize is removing the need to use them at all.
From Ground Truth · Everyone Is Adopting AI ↗Agents are starting to sit between people and the companies they buy from. They read everything, and they only shortlist what they can verify.
From Ground Truth · When Your Customer Is No Longer Human ↗This is the AgentNetwork Framework, our open model of how an organisation works once people and agents share the work. Most firms start with the tools. The tools are the least of it.
Everyone rents the same intelligence. The difference is what your organisation knows, who decides, and how the work is run.
We think like advisors and prove it like builders. Everything here is something you can open, watch or use today.
A strategic approach to transforming a global marketing team: where its attention goes today, which work agents should take on, and how roles change so people spend their time on the judgement and the relationships. The details stay with the client.
We didn’t build them an AI tool. We wrote down how the firm works in a form an agent could reason about, and built a chief of staff agent on top, with people holding the relationships and the judgement.
Which decisions stay human, which are agent assisted, which are agent led with approval. The relationship turned out to matter more than the technology.
The brief was technical. The motivation was civic. Who AI ends up serving is the most important question in the field.
Agents take tasks. People keep the role: the judgement, the relationships, the accountability. Separate the two and most of the fear about AI goes with it.

The breakthrough in 2017 came from taking machinery away. The same is true inside a business. The best agent is often the one that makes a process unnecessary.
“Playing simple football is the hardest thing there is.”Johan Cruyff · on screen in Tokyo

With infinite capability, the hardest question is what to stop doing. Most organisations have years of it.
From Ground Truth · The Work We Should Stop Creating ↗Smaller teams, fewer systems and tighter budgets used to be the constraint. In this shift they’re the head start. There’s less to unwind before the new way of working can begin.

Not human against agent, and not human replaced by agent. People at the centre, with agents handling the work below the threshold of judgement.
A short, senior engagement inside the operating reality. Where attention goes today, where it should go, and what agents should do first.
We prove the strategy by building it, with the people who’ll run it. Nothing arrives as a recommendation that could have arrived as a demonstration.
We stay while the organisation learns to run what we built together, then hand it over properly. Done with people, not to them.
Strategy and delivery in the same room. Decades of building at Microsoft, Nokia, UEFA, Art Basel and Minecraft, and a generation native to the tools.
Twenty five years at the frontier of digital: Microsoft, Nokia, UEFA, Art Basel. Perplexity Fellow. Writes Ground Truth. Brings the strategy, the technical depth and the senior advisory practice.
Operational leadership across Microsoft, Nokia and Minecraft Studios. Brings the delivery model, the financial discipline and the network of builders.
Builds the agents the firm runs on. Part of the first generation of operators native to these tools.
Craig’s essays on the intelligence era and what we choose to do with it. The writing does in public what the studio does in private: it pays attention.
New here · Start with the guide ↗Tell us what’s taking your organisation’s attention. You’ll get a first read straight back, and Craig or Fergus will reply to you personally.
Not ready to talk? Run the Attention Audit or map one workflow first.
The decision we keep delaying is how we work.