Judgement: The skill AI can't replace
Free masterclass page 3 of 3
Your judgement toolkit
Three organisations. Three industries. Three different constellations of traps. The scale was large but the mechanisms were not. Overconfidence operates the same way in a team of eight as it does in a surgical ward. Groupthink is no less potent in a Monday morning stand-up than in a Pixar screening room. And incentive misalignment doesn't need a stack ranking system to do its damage. A single misaligned bonus target will do.
The same underlying logic holds at every scale: no single tool fixed the problem. A portfolio of practices, applied across the layers where judgement actually operates, changed the conditions in which decisions were made.
Takeaway #6: Science-backed tools boost judgement
Scientists are clear on this: there is no one-to-one mapping of trap to cure for the errors that involve automatic associations or motivated reasoning. Overconfidence, for example, is partially addressed by tracking your predictions over time, partially by searching for reasons you might be wrong, partially by ensuring people in a group commit to their views independently before discussing them. None of these alone solves it. Together, they build the kind of thinker, and the kind of team, that overconfidence finds harder to take hold in.
The same is true for virtually every deep trap. Each is addressed by several tools simultaneously. Each tool addresses several traps.
Here are seven such science-backed tools, organised across the three layers where your judgement operates.
Takeaway #7: AI strengthens judgement when you bring the discipline
AI can be an extraordinary thinking partner. It can surface base rates you wouldn't have found, generate counterarguments you wouldn't have thought of, and spot patterns in your reasoning you didn't notice. But it responds to what you give it. If your process is good, if you're asking the right questions, giving it full context, and challenging what it produces, AI makes your judgement faster and more thorough. If your process is flawed, if you're asking leading questions, withholding context, or accepting the first polished output without scrutiny, AI accelerates the flaw.
The discipline is the same one this entire masterclass has been about: checking your frame, seeking disconfirmation, building structure around the moments where judgement is most at risk. AI doesn't change what good judgement requires. It raises the stakes for practising it. If your process is good, AI makes it faster and more thorough. If the process is flawed, AI makes it worse with more confidence, more polish, and more speed.
The tools come first. AI is the accelerant, not the engine.
Each of the seven tools you just read works with AI. Reframing: ask AI to generate alternative problem definitions you would not have considered. The outside view: ask it for base rates and reference classes. Flipping sides: ask AI to build the strongest case against your preferred position. If you find yourself reluctant to do that, or reluctant to take the result seriously, notice the reluctance. It's telling you something about how committed you are to the conclusion before you've tested it. Structured elicitation: have team members consult AI independently, with different questions, before the meeting. The risk is also real: team members consulting the same AI with the same query before a meeting destroys the independence the tool was designed to protect.
The risk isn't AI itself. It's what you stop doing because AI is there. If you use it only to produce answers faster without using it to challenge those answers, the judgement muscle that does the challenging quietly weakens. The tools in this masterclass exist to keep that muscle working.
Takeaway #8: Tools can boost entire processes
These tools, and more, can be employed individually as needed. Or, better yet be embedded in your step-by-step process or protocol that you follow for making a specific decision. You can create a portfolio of protocols for your most frequent or high-impact decisions. Decisions like:
-
What gets your attention? (prioritisation under competing demands)
-
Who do you bring on the team? (hiring)
-
Is the problem big enough to report up the ladder? (escalation)
-
How do you handle a difficult client or colleague situation? (communication, disagreement, influence and so on)
-
Which prospects should you invest time in chasing? (new client pursuit)
Go deeper on your own with The Judgement Toolkit: Starting with the decisions managers face most often, it includes step-by-step protocols for some of the toughest judgement calls you have to make. Each protocol has the right tools built in, so you follow the process and the science takes care of itself. Behind the protocols sit twelve tool families in full depth, corresponding decision traps, and an implementation guide built on the behavioural science you'll read about in the next section, so the tools don't stay on the page.
Access the Judgement Toolkit →
Making it stick
You now have a sharper picture of how your judgement works and where it's vulnerable. You've seen what happens when organisations change the conditions around their decisions. You have tools.
None of it matters if you don't use them next Tuesday, when you're under pressure, running late, and the meeting has already started.
That's the gap most training never closes. The science of behavioural change does.
That's the final piece of this masterclass: how can you help yourself and your team to actually use these tools, consistently, under the real pressures of daily work?
Takeaway #9: Behavioural science makes change stick
Better judgement sticks not through willpower or vigilance but through design: the way the tool is introduced, the way your environment is set up, the way your routines run, the way you interact with others when decisions are being made.
What follows uses the EAST framework: make it Easy, Attractive, Social, Timely. Created by the UK's Behavioural Insights Team, applied across domains from public health to financial planning, and adapted here for embedding decision tools in daily work.
Before you look at these practices, take a moment to think about one tool from this masterclass that you would like you and your team to adopt. What do you think will get in the way? Be honest.
It is usually one of five things: not sure exactly what to do (clarity), it feels like too much effort (friction), can’t see what it would fix for you specifically (attractive), nobody else does it (social), it will quietly disappear after a busy week (timing). Whichever one you recognise, start with the practices in that category.
Each of those barriers has specific, evidence-based strategies designed to remove it. The Judgement Toolkit includes eight, each one designed to work with any tool in this masterclass.
But you don't need the full set to start. You can begin with the simplest version of each principle right now:
Make it Easy. Remove one step of friction between you and the tool. If it's structured elicitation, put the template in the calendar invite so it's there before the meeting starts.
Make it Attractive. Connect the tool to your most recent painful decision. Not "better decisions in general" but "that hire last quarter that went wrong." Specificity creates motivation that abstraction can't.
Make it Social. Use the tool visibly, once, in front of your team. Not as a mandate. As a demonstration. If one person does it, it's a quirk. If the leader does it, it's a signal.
Make it Timely. Attach the tool to a specific recurring moment. "Before every Monday strategy meeting" is a trigger. "When I remember" is not.
That's enough to begin. One tool, one barrier, one principle. Give it two weeks before adding anything else.
There is a temptation, having seen all of this, to try to change everything at once. Resist it. Habits form slowly, far more slowly than most people expect, and attempting to build several simultaneously almost always means none of them take hold. Look at your diagnostic profile. Look at the barrier you named earlier. Find the practices that address your most important problem, the one that will get in your way tomorrow. Start there. The goal is not to use every tool and practice in this masterclass all at once. It's to become the kind of decision-maker, and the kind of team, that uses the right tools reliably, under pressure, when it counts.
Before you go
You've just spent an hour looking honestly at how your judgement works. Most professionals never do this deliberately, and fewer still do it at a moment when the tools around them are changing faster than at any point in their careers.
AI will keep getting more capable. It will take on more of the analysis, more of the processing, more of the pattern recognition. That's not a threat to your role. It's a sharpening of it. What remains, the work that will define whether you and your team make good calls or merely efficient ones, is judgement. And judgement, unlike processing power, does not improve automatically. It improves by design.
What you've done here is what a decision expert would do first: look honestly at how judgement works, identify where it breaks, and find the tools that address it. If you had someone like that embedded in your team, they'd observe how decisions actually get made, guide you out of traps you can't see from the inside, co-build protocols around the decisions that matter most, and help you design an environment where good judgement is the default rather than the exception.
You don't need to do all of that alone. The Judgement Toolkit is the closest thing to having that expert on your desk: step-by-step protocols for the decisions you face most often, the tools built into every step, and a behavioural change guide for making them stick. It's where most people go next.
Have any questions? Let's talk →
A small ask. I'd like to know what landed and what didn't. Give feedback →
Thank you for being here.