Philosophy
Whose judgement do we improve?
It’s the question we measure our work by, and we answer it at two scales. The larger is the decisions that reach thousands of lives: policy, and the not-for-profits, peak bodies and advocacy organisations we work for, which shape it. The smaller is the person at their desk, with the room and the confidence to judge well.
The motive
Why we work with for-purpose organisations.
For-purpose organisations, not-for-profits, community services, peak bodies and advocacy organisations, carry a community’s voice into policy, often with very few staff. They come first for us, and the reason is impact. Give a small organisation the capacity of a much bigger one, and the gain lands on the people it serves. That is the work we want to be doing: it fits our moral compass, and it uses the full breadth of our expertise, from the science of judgement and learning to the systems and training that turn it into capacity on an ordinary working day.
The climb
Impact is the measure.
The reach that matters to us is how high up the decision chain the ideas travel, more than fame or the size of an audience. Reaching one person who writes national policy can outweigh a million podcast downloads.
Advocacy organisations are the way up that chain. Politicians and senior officials rely on them to say what the problems are and what the evidence supports. Those organisations are small and stretched, and they rarely have the staff to weigh evidence or turn research into argument. So the quality of policy depends, in part, on the quality of advocacy. Helping them work better is the most direct route we know from good thinking to better decisions for the people policy reaches.
A climb in four levels. At the bottom, a person at their desk; above, an organisation; above that, a sector of peak bodies and advocacy organisations; at the top, the people who write policy. Tailored Ideas works at the first three levels. The sector reaches the top level through submissions, briefings and evidence, where one better judgement reaches thousands of lives.
What we’re for
Carrying the science across.
Cognitive science, the study of how people reason, learn, judge and decide, knows a great deal about how people do their best work. Yet much of it stays locked away in research papers and expert heads, and never reaches the working lives it could improve. Applied cognitive science is the craft of carrying it across: turning what is known about the mind into something a person can use on an ordinary working day. Today that craft is landing in artificial intelligence, the change reshaping how almost everyone works. AI is the setting, not the destination.
Two halves, then, one for each scale in the question at the top:
One craft, carrying the science of how people think, learn and decide into everyday work, over two halves. Judgement, at the scale of decisions that reach thousands of lives, moves decisions toward what holds up and away from the answer that only sounds right, and climbs from organisation to sector to policy. Expansion, at the scale of the person at their desk, moves work toward what stretches without overwhelming, away from grind and overload.
The craft · carrying the science of how people think, learn and decide into everyday work
Climbs: organisation, sector, policy
Judgement
The scale: decisions that reach thousands of lives
Toward: decisions made on what holds up
Against: the answer that only sounds right
Expansion
The scale: the person at their desk
Toward: work that stretches without overwhelming
Against: grind, and overload
Judgement: decisions that reach thousands of lives.
Decisions made on what holds up.
Many consequential decisions, in policy, in workplace culture and in the systems people work inside, are made on intuition that was never checked. Well-earned expert intuition is powerful. The trouble is the alluring intuition: the answer that sounds right because it is compelling rather than correct.
Help the people behind a submission, a board decision or a program judge better, and the change reaches everyone those decisions touch. It’s the subject of Ryan’s research in cognitive science (PhD, University of Queensland): how people judge and decide, and how they tell real insight from what only sounds profound.
Isn’t intuition a good thing?
Often, yes: expert intuition is powerful when it has been earned, from plenty of experience and honest feedback on whether you were right. Where that feedback is missing, an answer can feel like expertise when it is only fluent.
How we work on that with leaders and boardsA two-by-two grid comparing whether an answer sounds right with whether it is right. Sounds right and is right: well-earned intuition. Sounds wrong and is right: the uncomfortable truth. Sounds wrong and is wrong: the easy no. Sounds right but is wrong: the alluring intuition, highlighted as the one to catch.
Where the judgement half shows up in our work:
In the arguments organisations make. Submissions and briefings built on the evidence of what persuades and what holds up, with every claim checked against its source.
The judgement science behind sound submissionsIn the systems we build. Fluent writing feels accurate, so the first plausible draft tends to get waved through. Every draft our systems produce arrives with a prompt on what its reviewer should check: the prompted review.
How the prompted review worksAt the top table. Boards and senior leaders are often cocooned from honest feedback. We build the knowledge and the honest feedback that good intuition depends on, rather than asking people to think harder.
Judgement for leaders and boards
Expansion: the person at their desk.
Better tools should make for better workdays.
Around
80,000
hours of your life at work
Time on that scale should give something back, not just be time to get through.
Why does that band matter?
People are most engaged when a task sits just beyond their current ability: hard enough to demand focus, not so hard it overwhelms. Below that line, work turns into grind, the repetitive parts that leave you running on autopilot. Too far above it, pressure tips into anxiety. Between the two lies a narrow band where attention sharpens, skill grows, and the hours feel worth something.
This is where working well with AI can take people. Taking the grind off their plate frees more of their time for that growing edge, and lifting what they can do brings work within reach that was once beyond them. Used this way, AI can be a catalyst for joy, mastery, and meaning.
A more human way to work.
Used well, AI takes the grind off people’s plates and lifts what they can do, and one large study found the gains reached past output, to how people worked and whether they stayed. The reason runs deep. People are most engaged when a task sits just beyond their current ability, in a narrow band between boredom and anxiety, and that is where working well with AI can take them.
What did the study find?
Output
15% more issues resolved per hour, on average. Around 30% for less experienced and lower-skilled workers.
Capability
When the AI system went down, workers who had used it still handled conversations faster than their own pre-AI baseline.
Experience and retention
Requests to escalate to a manager fell by roughly 25%. Attrition among agents with less than six months’ tenure fell by around 10 percentage points (a result the authors treat with more caution).
Brynjolfsson, Li and Raymond, Generative AI at Work, The Quarterly Journal of Economics, 2025: 5,172 customer-support agents at one Fortune 500 firm. The authors are clear that it covers one tool, in one firm, in one occupation.
It’s a strong study: thousands of workers, a staggered rollout, and the whole chain from output to retention in one dataset. It’s also one AI tool, in one firm, in one occupation, and its authors say so: the findings shouldn’t be stretched to every job and every system. They treat the drop in attrition with more caution than their main results, because a person leaves a job only once, which makes that result harder to model. We cite it because it’s the best single source on the question, and we hold it to the strength it carries.
The imagination problem.
If the opportunity is this clear, why does it stall? Rarely for lack of access or technique. People stay with the familiar, and most can’t yet picture AI inside the judgement-heavy parts of their work, where it would matter most. What changes that is seeing it done, well, on a real task from their own desk.
Why does it stall?
A survey of 767 knowledge workers. Berman and colleagues, Irrational Labs, The AI Workplace: New Research on Employee Adoption, 2025. The authors call it an intention-action gap.
Many organisations have done what looks right: bought the tools, granted access, written the policies. Still the gains stall, with most people using AI for the obvious tasks while its real potential sits untouched.
The barrier is rarely access or technique; both can be solved. It’s psychological. When a new way of working appears, even an obviously better one, people tend to stay with the familiar. Researchers call this status quo bias, and at its root is loss aversion: we feel what we might lose far more sharply than what we might gain.
At work, what people most fear losing is time. Between back-to-back meetings, hours spent learning a new tool feel like a loss right now, while the time it might give back stays abstract and far off. So the familiar wins, not because people weighed it up and chose it, but because the scales were tipped before they looked.
Like most cognitive biases, it works largely beneath awareness. People aren’t refusing the opportunity; they don’t see it, so there’s nothing to choose for or against.
With AI, this shows up as a failure of imagination. Most people have used it to draft an email or summarise a document, tasks that look like old tools doing familiar jobs. Far fewer can yet picture it inside the judgement-heavy parts of their work, where it would matter most. Until they can, no amount of prompt cards or tool tutorials will move them.
In the same survey, the strongest predictor of use was managerial support: 79% used AI where their managers endorsed it, 34% where they didn’t. And 87% knew at least one person who used AI; that group was three times more likely to have used it themselves in the past week. Both findings fit the imagination problem: the possibility comes into view when people see it done, close to home.
The evidence on The ScienceSo how do you train for it?
Not platform fluency.
AI tools change on a timescale of months. Learn this month’s menus and shortcuts and you’ve bought a skill with an expiry date: by next quarter the buttons have moved and a new model has landed.
What lasts is the thinking underneath: knowing what these tools are good at and where they quietly fall down, telling a sharp answer from one that only sounds right, and spotting the tasks in your own work worth handing over. Build that, and people adapt to whatever the tools look like in a year, or two.
Instruction informs; intervention changes behaviour.
Because the barrier is psychological, being told the upside isn’t enough. You can show someone a slide about what AI could do for them, watch them nod along, and find that nothing has changed by Monday. What moves people is experiencing it for themselves: taking a real task from their own desk, the kind they have been quietly dreading, and watching it get done, and done well, in front of them. That’s when a new way of working takes hold.
Underneath it all
It all runs on the science of the mind.
Our values are our own. Our methods come from one place: cognitive science, the study of how minds work. The science of learning and decision psychology are views from within it; measurement and applied AI are how it gets to work. Together they are how we answer the question at the top of this page, at both scales.
The umbrella discipline
Cognitive science
The study of how minds work.
The science of learning
How people build durable skill.
Decision psychology
How people judge, decide, and resist or accept change.
Measurement
How to know what changed.
Applied AI
What the tools can do for how people think, and how that’s shifting.
None of these is enough in isolation. The integration is the point.
Start with a conversation.
Bring what you’re trying to change, whether it’s the case you put to government or the way your team works.