Why Curiosity Beats Shame in Software Retrospectives

There’s a moment in therapy that therapists call ‘the shift’—when you stop drowning in your patterns and start watching them with fascination. You realise you’ve been having the same argument with your partner for three years, and instead of feeling like a broken record, you start laughing. ‘Oh, there I go again, catastrophising about the dishes.’ The pattern doesn’t vanish overnight, but something fundamental changes: you’re no longer at war with yourself.

What if software teams could experience this same shift?

The Drama We Know By Heart

Every team has their recurring drama. Maybe it’s the sprint planning meeting that always runs two hours over because nobody can agree on story points. Perhaps it’s the deployment Friday that inevitably becomes deployment Monday because ‘just one small thing’ broke. Or the code review discussions that spiral into philosophical debates about variable naming or coding standards more generally, whilst the actual logic bugs slip through unnoticed.

We know these patterns intimately. We’ve lived them dozens of times. Yet most teams approach retrospectives like a tribunal, armed with post-its and grim determination to ‘fix our dysfunction once and for all.’ We dissect our failures with the energy of surgeons operating on ourselves, convinced that enough shame and analysis will finally make us different people.

But what if we’re approaching this backwards?

The Mice Would Find Us Fascinating

Douglas Adams had it right when he suggested that mice might be the truly intelligent beings, observing human behaviour with scientific curiosity. Imagine if we could watch our team dynamics the way those hyperintelligent mice observe us—with detached fascination rather than existential dread.

‘Interesting,’ the mice might note. ‘When the humans feel time pressure, they consistently skip the testing phase, then spend three times longer fixing the resulting problems. They repeat this behaviour with remarkable consistency, despite claiming to have “learned their lesson” each time.’

The mice wouldn’t judge us. They’d simply observe the pattern, maybe take some notes, perhaps adjust their experiment parameters. They wouldn’t waste energy being disappointed in human nature.

The Science of Predictable Irrationality

Behavioural economists like Dan Ariely have spent decades documenting how humans make decisions in ways that are wildly irrational but remarkably consistent. We’re predictably bad at estimating time, systematically overconfident in our abilities, and reliably influenced by factors we don’t even notice. These aren’t bugs in human cognition—they’re features that served us well in evolutionary contexts but create interesting challenges in modern day work environments.

Software teams exhibit these same patterns at scale. We consistently underestimate complex tasks (planning fallacy), overvalue our current approach versus alternatives (status quo bias), and make decisions based on whoever spoke last in the meeting (recency effect). The beautiful thing is that once you name these patterns, they become less mysterious and more laughable.

Curiosity as a Debugging Tool

When we approach our team patterns with curiosity instead of judgement, something magical happens. The defensive walls come down. Instead of ‘Why do we always screw this up?’ we start asking ‘What conditions reliably create this outcome?’

This shift from shame to science transforms retrospectives from group therapy sessions into collaborative debugging. We’re not broken systems that need fixing—we’re complex systems exhibiting predictable behaviours under certain conditions. Complex systems can be better understood through observation, and sometimes influenced through small experiments, though the outcomes are often unpredictable

Consider the team that always underestimates their stories. The shame-based approach produces familiar results: ‘We need to be more realistic about our estimates.’ (Spoiler alert: they won’t be.) The curiosity-based approach asks different questions: ‘What happens right before we make these optimistic estimates? What information are we missing? What incentives and other factors are shaping our behaviour?’

The Hilariously Predictable Humans

Once you start looking for patterns with curiosity, they become almost endearing. The senior developer who always says ‘this should be quick’ right before disappearing into a three-day rabbit hole. The product manager who swears this feature is ‘simple’ whilst gesturing vaguely at convoluted requirements that would make a vicar weep. The team that collectively suffers from meeting amnesia, forgetting everything discussed five seconds after the meeting ends.

These aren’t character flaws to be eliminated. They’re what Dan Ariely would call ‘predictably irrational’ behaviours—systematic quirks in how humans process information and make decisions. The senior developer genuinely believes it will be quick because they’re anchored on the happy path scenario (classic anchoring bias). The product manager sees simplicity because they’re viewing it through the lens of user experience, not implementation complexity (curse of knowledge in reverse). The team forgets meeting details because our brains are optimised for pattern recognition, not information retention across context switches.

We’re not broken. We’re just predictably, irrationally human.

Practical Curiosity: Retrospective Questions That Transform

Instead of ‘What went wrong this sprint?’ you might like to try:

‘What hilariously predictable human things did we do again?’‘If we were studying ourselves from the outside, what would be fascinating about our behaviour?’‘What patterns are we executing so consistently that we could almost set our watches by them?’‘Under what conditions do we make our most questionable decisions?’What shared assumptions inevitably led to this sprint’s outcomes?‘What would the mice find interesting about how we work?’

These questions invite observation rather than judgement. They make space for laughter, which is the enemy of shame. And they reduce the role of shame—the antithesis of learning.

The Liberation of Accepting Our Programming

Here’s the paradox: accepting our patterns makes them easier to change. When we stop fighting our humanity and start working with it, we find leverage points we never noticed before.

The team that always underestimates might not become perfect estimators, but they can build buffers into their process (Cf. TOC). The developer who disappears into rabbit holes can set timers and check-in points (such as Pomodoros). The product manager can be paired with someone who thinks in implementation terms.

We don’t have to become different people. We just have to become people who understand ourselves better.

AI as a Curiosity Amplifier

Here’s where artificial intelligence might genuinely help—not as a problem-solver, but as a curiosity amplifier. AI excels at exactly the kind of pattern recognition that’s hard for humans trapped inside their own systems.

Pattern Recognition Beyond Human Limits

AI could spot correlations across longer timeframes than teams naturally track. Perhaps story underestimation always happens more, or less, after certain types of client calls, or when specific team members are on holiday. Maybe over-architecting solutions correlates with unclear requirements, or planning meetings grow longer when the previous sprint’s velocity dropped.

These are the kinds of subtle, multi-factor patterns that human memory and attention struggle with, but that could reveal fascinating insights about team behaviour.

Systematic Curiosity Drilling

More intriguingly, AI could help teams ask better layered questions: ‘We always over-architect when requirements are vague → What specific types of vagueness trigger this? → What makes unclear requirements feel threatening? → What would need to change to make simple solutions feel safe when requirements are evolving?’

This is the kind of systematic curiosity that therapists use—moving from ‘this is problematic’ to ‘this is interesting, let’s understand the deep logic.’ AI could be brilliant at sustaining that investigation without getting distracted or defensive.

The Crucial Cautions

But here’s what AI absolutely cannot do: the therapeutic shift itself. The moment of laughing at your patterns instead of being tormented by them? That’s irreplaceably human. AI risks creating surveillance anxiety—the sense that someone (or something) is always watching and judging.

There’s also the fundamental risk of reinforcing the very ‘fix the humans’ mentality this approach seeks to avoid. AI pattern recognition could easily slide back into ‘here are your dysfunctions, now optimise them away.’

The sweet spot might be AI as a very patient, non-judgmental research assistant—helping teams investigate their own behaviour more thoroughly. The humans still have to do the laughing, the accepting, and the choosing. But AI could make the curiosity richer and more evidential.

Just remember: the mice observed the humans with detached fascination, not with algorithms for improvement.

The Recursive Gift

The most beautiful part of this approach is that it’s recursive. Once your team learns to observe its patterns with curiosity, you’ll start applying this same gentle scrutiny to your retrospectives themselves. You’ll notice when you slip back into judgement mode and laugh about it. You’ll develop patterns for catching patterns.

You’ll become a team that’s as interested in how you think as in what you build. And that might be the most valuable code you ever debug.

The Pattern That Doesn’t Disappear

Your recurring drama won’t vanish. The sprint planning will probably still run long sometimes. The ‘quick fix’ will occasionally become a weekend project. But your relationship to these patterns will transform. You’ll work on them without the crushing weight of believing you should be different than you are.

And in that space—between pattern and judgement, between observation and criticism—you’ll find something remarkable: the room to actually change.

The mice would be proud.

Further Reading

Adams, D. (1979). The hitchhiker’s guide to the galaxy. Harmony Books.

Ariely, D. (2008). Predictably irrational: The hidden forces that shape our decisions. Harper.

Netó, D., Oliveira, J., Lopes, P., & Machado, P. P. (2024). Therapist self-awareness and perception of actual performance: The effects of listening to one recorded session. Research in Psychotherapy: Psychopathology, Process and Outcome, 27(1), 722. https://doi.org/10.4081/ripppo.2024.722

Williams, E. N. (2008). A psychotherapy researcher’s perspective on therapist self-awareness and self-focused attention after a decade of research. Psychotherapy Research, 18(2), 139-146.