If you don't ask where you are fragile, the next shock will answer for you.

Companies don't manufacture fragility out of stupidity. They manufacture it through behaviour that worked for years: capacity at full, one proven procedure, exceptions through the boss. In an exponential world, that breaks on the first big shock.

Four levels: from fragile to antifragile

Same company, same people, same shock. The only difference is where decision-making sits in the system. Which level is your company, division or team on today?

Level 1

Fragile system

Micromanagement, the leader as dispatcher
The boss is on holiday. Nothing gets put out until they are back.

All the intelligence sits in a single node. Every step goes through the leader. A small load or their absence is enough to bring operations to a stop.

Productivity~15 %
Leader's stress95 %
Level 2

Robust system

Rigid rules and directives
We may call the fire brigade after three signatures. We survive a small shock, a bigger one destroys us.

The intelligence is in rules and tables. People follow the process even where it makes no sense. A change the rules did not anticipate makes the system jam.

Productivity~30 %
Leader's stress75 %
Level 3

Adaptive system

Autonomy and local agreements
People put it out themselves. Then everything is restored to how it was.

Teams agree among themselves and can take a time-out. They survive the shock and return to their original performance. Tired and no further ahead.

Productivity~50 %
Leader's stress60 %
Level 4

Antifragile system

The leader as architect of the environment
We run drills on purpose. And we rebuild the system based on what they show.

The leader does not manage people, but the environment. Under controlled pressure the team improves its own system and grows with every challenge. Certainties are removed deliberately, not by accident.

Productivity~95 %
Leader's stress5 %
Fragility: high stress, low outputAntifragility: low stress, growth from stress

Productivity and leader's stress are rough estimates from the course of the Living Crossroads simulation. They show a direction, not a benchmark for your company.

You cannot get this from a picture. That is why we built a simulation in which leaders experience it.

How Living Crossroads works

Living Crossroads: one day when your leaders experience it in their own bodies

  • 16–50 leaders
  • 1 day
  • A personal AI guide for everyone

Under pressure we do more of what made us successful: more control, more directives, more reports. Every push without a change of the environment blocks another crossroads. The system does not collapse because of the people, but because of its design.

In Living Crossroads people physically step into the role of elements of a system and go through all four levels of control. Through movement, emotion and real stress. Nobody has to explain to them why more rules do not help. They see it in the performance of their own team.

Reading about it is not enough. Talking about it is not enough. You have to experience it.

Participants of Living Crossroads on a lawn between cones, one of them standing on crates and directing the others
Level 1 in practice: a single dispatcher runs the whole crossroads.

  • The end of managerial overload

    The E-A-D rule: eliminate, automate, delegate. How to stop being the bottleneck of your own operations.

  • Clearing the bureaucratic debt

    The eight most common forms of waste in a company: work standing still, work piling up, people working below their potential. And what can be cut right away.

  • AI for an operational jump

    How AI tools cut routine work from days to hours and open up capabilities that used to belong to a specialist.

  • One step and an AI guide

    Everyone leaves with one thing they will stop or automate tomorrow. And with a personal AI assistant that keeps them at it.

Why it changes how a unit works: a physical experience leaves an emotional trace that no presentation can create. And within a few hours it breaks the illusion that more work and more rules mean a better result.

Who it is for: leadership and the B-1 level, directors of large processes and regulated units such as risk, security or controlling. For the people who will genuinely back a change, or stop it.

What leaders from banking took away

“We have a lot of regulated processes, and inside them you assume you are doing the best that can be done. This experience shows you in your own body that there is a solution you have not even thought of yet — and that what we usually lack is not the competence, but the courage to try it. What I got most from was realising that linear imagination is no longer enough, we need an exponential one.”

Who he would recommend it to: “Anyone who designs processes with guardrails. I would like to bring it to my whole team.”

Matej ŠtalmachDirector, Retail CRM, Tatra banka
“I was genuinely shocked by the difference in productivity between the beginning and the end. When I looked at the initial and the final state, it was a fundamental, multi-fold improvement — even after I account for the risks. I value stability extremely highly myself, and that is exactly why the exercise forced me to open up to the question of how radically even a completely established process can be changed.”

Who she would recommend it to: “It makes the most sense for people who genuinely influence change — who can support it or stop it. Ideal for control and regulated units, so that they learn to think in these terms as well.”

Katarína RolnáDirector, Security and BCM, Tatra banka
“It is one of the best things we have brought into the programme, because it shows how we can be — completely unconsciously — resistant to change. Dropping a bad habit is easy, but being willing to give up something that works, that is real unlearning.”
Peter PápayHead of Learning and Development, programme owner, Tatra banka
Two more references
“The strongest part was that you live it through movement — that is a completely different experience from theory. The whole time I was thinking: what kind of crisis do I have to trigger for us to reconsider the things we treat as set in concrete today? It is like cold exposure training — you change your relationship to the cold: suddenly it is not the thing that makes you ill, but the thing that makes you stronger. That is exactly how the attitude to change shifts.”

Who he would recommend it to: “Anyone leading a change — and also the people who actually carry it out, because after an experience like this, change is far easier to sell.”

Peter ŠtillaRegional director RB West, Raiffeisen banka
“I realised that when you really want to change something, you need a big enough crisis — otherwise you always slip back into the old ways. And it is a powerful tool when you need people to buy in: you do not have to order anything, they understand for themselves why some things have to be done completely differently.”

Who he would recommend it to: “For the B-1 level and anyone responsible for a large process — in business, risk, operations or finance.”

Stanislav ŠimoDirector, Controlling, Tatra banka

We adapt the scope, format and number of participants to your leaders.Let us discuss it over coffee. No slides, no commitment.

I want Living Crossroads for our leaders

The catalogue of fragility sins

A sin of fragility is an inherited virtue: leader behaviour that was the best thing to do in a stable world, and that was rightly rewarded for years. Every such reflex creates a counterfeit of resilience in the company. Something you point at as proof of readiness, and that does not hold under a real shock.

Do not read it as a list of mistakes. For every sin you will find why it made sense, how to spot it in yourself and one thing for Monday: within an hour, no budget, no approvals. We give numbers only where there is a source.

The seven from the stage

These seven were in the talk.

1
A question from the start of the talk

Zero slack means order

Free capacity gets filled with work right away.

Counterfeit of resilience: “When the peak comes, people will pull it off.”Capacity slack disappears
More detailsHide details

In a stable world

What I doI plan the capacity of key people to the full. An open slot in the calendar reads as a planning error, so I fill it immediately. Slack has no line of its own in the budget, so every round of cuts removes it together with the waste.
Why I do itI believe an unused hour is an hour paid for and wasted.
Why it made senseWhen work arrives predictably, free capacity is a pure cost. Higher utilisation means higher throughput on the same fixed costs. Queues have time to dissolve, because nothing arrives in bursts.

In an exponential world

What happens under a shockWaiting time does not grow evenly with utilisation, it grows steeply. The queue factor is 1 at 50 % utilisation, 4 at 80 %, 9 at 90 % and 19 at 95 % (hard research: Kingman's formula, 1961). When a shock arrives, there is nothing left to cover it. Capacity is taken from live operations and the system starts to cannibalise itself.
Counterfeit of resilienceOvertime as the reserve. „When the peak comes, people will pull it off.“ The shock absorber you point at has long been used up.
How to spot it in yourselfTake the last three requests your key team handled. For each, compare the time from request to delivery with the time someone actually worked on it. If the work spent most of that time waiting, your performance is run by the queue, not by people. Kingman's formula explains why: as utilisation grows, waiting does not increase evenly but steeply (hard research: Kingman).
New beliefI believe free capacity is a purchased ability to respond to a shock.
New behaviourTeams that carry unpredictable work I plan at 70–80 % (opinion: Reinertsen, The Principles of Product Development Flow, 2009). Slack is a named item in the capacity model, not what happens to be left over.
What it creates in the companyA shock absorber nobody has to approve when a shock arrives. The team absorbs the swing without taking capacity out of live operations, and in calmer periods it has time to improve.
Where to start on Monday: Open the capacity plan or the calendar of a key team for next month and count how many hours are not assigned to anything. Put that number at the top of the next team meeting.

When it still holds: Where work demonstrably arrives evenly and the queue is measured: manufacturing with a level takt, standardised transactions. Condition: if no queue longer than one takt appeared in a year, full utilisation is the right call.

2

Firefighters as the elite

The rescue is rewarded, not the problem ceasing to occur.

Counterfeit of resilience: “We have people who always pull it off.”Capacity slack disappears
More detailsHide details

In a stable world

What I doPromotions, praise and visibility go to the people who saved a burning project. The person who spent three months removing the cause so that nothing would burn is never mentioned in the meeting. They have no result to point at.
Why I do itI believe the most valuable person is the one who saves whatever is on fire.
Why it made senseWhen failures are rare and local, a fast repair is the cheapest answer. Investment in prevention pays back only after many cycles, and with rare failures perhaps never. Rewarding fast fixers is then the right allocation.

In an exponential world

What happens under a shockFirefighting consumes the time that would go into removing causes, so the fires keep coming. The organisation rewards last-minute solutions more than prevention, and the senior layer gradually fills with „war heroes“ who pick people like themselves (hard research: Repenning & Sterman, California Management Review, 2001). The company then knows how to repair, but not how to rebuild.
Counterfeit of resilienceThe reliable rescuer. „We have people who always pull it off.“ Resilience rests on names, not on the system.
How to spot it in yourselfGo through the last ten promotions and exceptional bonuses in your unit. How many went for a visible rescue and how many for a problem that stopped recurring? Write down both numbers.
New beliefI believe the most valuable person is the one after whom things stop burning. And that they will get no praise for it until I introduce it.
New behaviourI introduce a second kind of recognition: for a problem that stopped recurring this year. A fixed share of the team's capacity goes into removing causes, and I measure it by a drop in incidents, not by the number of tickets closed.
What it creates in the companyPeople who have a reason to remove causes. The number of repeating incidents falls and the time freed from firefighting goes into further improvement. The loop described by Repenning and Sterman turns the other way.
Where to start on Monday: At your next team meeting, thank one person by name for a problem that stopped recurring thanks to them. If you cannot find anyone, that is a result too.

When it still holds: In the acute phase of a real crisis. And wherever the number of incidents of the same kind is not growing year on year — prevention there has no demonstrable return.

3

Title decides, not information

In a crisis the highest title decides, not the person who knows most.

Counterfeit of resilience: “We have a crisis team.”Decision slack disappears
More detailsHide details

In a stable world

What I doWhen a situation gets out of hand, the highest title in the room takes over the decision. Nobody ordered it, it happens by itself. The person who knows most about the problem speaks as an advisor, not as the one deciding.
Why I do itI believe that in an unclear situation the decision belongs to whoever is accountable for it.
Why it made senseAuthority reliably ends a dispute and saves time. When problems repeat, the most senior person has seen them most often over their career. Title is then a good proxy for knowledge.

In an exponential world

What happens under a shockIn a new situation, title stops correlating with knowledge. The decision is made by the person with the least detail, based on a retelling by the person with the most. What drops out of the retelling is exactly what the narrator did not consider important — and in a new situation that is often the crucial part. High-reliability organisations do the opposite and in a crisis migrate decisions to expertise regardless of hierarchy (hard research: Weick & Sutcliffe, Managing the Unexpected).
Counterfeit of resilienceA crisis team made up of the leadership. „We have a crisis team.“ It seats the highest titles, not the people who know the system.
How to spot it in yourselfRecall the last three crises or serious incidents. For each, write two names: who decided and who knew most about the problem. In how many of the three were they two different names?
New beliefI believe that in an unclear situation the decision belongs to whoever knows most about it.
New behaviourFor the most likely types of crisis I decide in advance who makes the call by type of situation, not by level. When a crisis hits, the most senior person hands the decision over out loud and stays in the room.
What it creates in the companyCrisis decisions that rest on the best available information and are not delayed by retelling. Leadership can focus on what belongs only to it: cost, reputation and external communication.
Where to start on Monday: Take the most likely type of crisis in your unit and write down the name of the person who should decide on the intervention. Send it to them and ask whether they agree.

When it still holds: For decisions whose impact reaches beyond the problem itself: reputation, legal risk, capital, communication with the regulator. The expert decides on the intervention, the title decides on its price and on what is said externally.

4
A question from the start of the talk

One best way

One procedure per job, the second one is not maintained.

Counterfeit of resilience: “We have ISO, everything is documented.”Slack of alternatives disappears
More detailsHide details

In a stable world

What I doFor every repeated job I find the best procedure and introduce it as the only one. I remove variants as a source of errors. A new person learns that one way and never hears about the others.
Why I do itI believe there is one best way to do every job and my role is to enforce it.
Why it made senseStandardisation lowers the error rate, shortens onboarding and allows growth without expertise having to grow with it. In a clear and stable environment it is literally the right answer.

In an exponential world

What happens under a shockWhat standardisation removed as variability was also a stock of alternatives. When conditions change, nobody knows another way. The boundary between the clear and the chaotic domain is not a slope but a cliff: you do not fall into the complicated, you fall straight into chaos (opinion: Snowden & Boone, A Leader's Framework for Decision Making, HBR 2007; domain descriptions, cynefin.io).
Counterfeit of resilienceA documented and certified process. „We have ISO, everything is documented.“ What is documented is the single path. Nobody can do the second one.
How to spot it in yourselfFor a process whose outage stops you: how many people can run it in a different way than the established one — and when did somebody last actually do it? Write down both numbers.
New beliefI believe a second way is an insurance policy I pay for in advance.
New behaviourFor processes whose outage stops the business, I keep at least one working alternative. And I have it run for real from time to time, not merely described.
What it creates in the companyA second path that people know from their own experience, not from documentation. When conditions change, the organisation has something to choose from and does not have to improvise under pressure.
Where to start on Monday: Pick one critical process and ask two people who run it: „If the main system or procedure went down tomorrow, how would you do it?“ Note whether they had an answer.

When it still holds: Where the environment is demonstrably stable and an error is expensive: transactions, safety, regulated operations, manufacturing to a fixed specification. Condition: you can name when the procedure stops holding, and somebody watches for it.

5

Deviation is an offence

Whoever departs from the process pays for it.

Counterfeit of resilience: “Zero deviations, a clean audit.”Learning slack disappears
More detailsHide details

In a stable world

What I doI treat every deviation from the approved procedure as a failure, regardless of the outcome. Someone who deviated and saved the situation is reprimanded just like someone who ruined it. An error ends in a search for the culprit and one more control.
Why I do itI believe a tolerated deviation is the beginning of decay.
Why it made senseWhen the procedure really is the best known one, every deviation is statistically the worse choice. Enforced compliance raises average performance. In a regulated environment it is a legal obligation on top of that.

In an exponential world

What happens under a shockA deviation is the first signal that the procedure is ceasing to hold — and punishment removes that signal first. Where errors are paid for, errors stop being reported. In a study of hospital teams, the better performing teams reported more errors, not fewer, because they did not hide them (hard research: Edmondson, Journal of Applied Behavioral Science, 1996; Edmondson, Administrative Science Quarterly, 1999). The company learns about the deterioration only from the result. And in a crisis for which no procedure exists, people trained on punishment freeze or ask upwards.
Counterfeit of resilienceA green compliance dashboard. „Zero deviations, a clean audit.“ Zero does not mean nothing happened. It means nothing is being reported.
How to spot it in yourselfHow many deviations from procedure or near misses were reported in your unit last quarter? Zero does not mean none occurred. In Edmondson's hospital study it was precisely the better performing teams that reported more errors (hard research: Edmondson 1999).
New beliefI believe a deviation is the cheapest signal that a process is ceasing to hold.
New behaviourI judge the decision separately from the outcome: first I ask what that person could see at that moment. I track the number of reported deviations and near misses as a good indicator. When it falls, that is a warning.
What it creates in the companyInformation that a process is ceasing to hold, at a moment when the fix is still cheap. And people who act in a situation without a procedure instead of freezing.
Where to start on Monday: For the last deviation you dealt with, ask the person who deviated what they could see at that moment and why they considered it right. Leave the assessment for afterwards.

When it still holds: Where the deviation itself breaks the law or endangers life. Even there, though, reporting is never punished — only an unreported workaround is.

6

All-in bet

Change as one big programme instead of a portfolio of small bets.

Counterfeit of resilience: “We have a plan and governance.”Slack of alternatives disappears
More detailsHide details

In a stable world

What I doI run change as one big programme: full scope, long horizon, one target state. A portfolio of small experiments reads to me as fragmentation. I decide once, at the start.
Why I do itI believe big change needs a big commitment, and that transformations fail because effort gets scattered.
Why it made senseConcentrating resources removes switching, brings economies of scale and clear accountability. When the future is predictable years ahead, one bet is cheaper than a portfolio.

In an exponential world

What happens under a shockThe decision was made at the moment you knew least. The longer the programme runs, the more you know and the more expensive it becomes to change direction. Of more than 16,000 large projects in Bent Flyvbjerg's database, only 0.5 % finish on budget, on time and with the promised benefit (hard research: Flyvbjerg & Gardner, How Big Things Get Done, 2023 — data from infrastructure and IT; the transfer to organisational change has not been measured directly). When a shock arrives, resources are locked in a programme half of which has no value yet.
Counterfeit of resilienceA transformation programme with a roadmap and a steering committee. „We have a plan and governance.“ The plan was made at the moment you knew least.
How to spot it in yourselfWhat share of this year's change budget is committed to initiatives that deliver their first measurable result more than a year out and cannot be stopped without writing off the whole investment? Write down the percentage. For comparison: of the large projects in Flyvbjerg's database, 0.5 % meet budget, schedule and benefit (hard research: Flyvbjerg & Gardner 2023).
New beliefI believe big change needs many cheap experiments with a limited downside.
New behaviourI protect the core. Part of the change budget goes into small bets with a capped loss and a result within months. Every bet has a pre-agreed point at which I stop it.
What it creates in the companyA portfolio that learns as it goes. Most bets fail cheaply, a few pay off. When a shock arrives, resources can be moved, because they are not locked in a single programme.
Where to start on Monday: For your largest running programme, write one sentence: „We will stop this programme if we do not see [measurable result] by [date].“ If you cannot write it, you know what the programme is missing.

When it still holds: When half a result has no value at all rather than half the value: construction, a core system migration, an acquisition. There the bet is sized from the actual outcomes of comparable projects, not from your own team's estimate.

7

The trusted supplier

Consolidation onto a single proven partner.

Counterfeit of resilience: “The supplier is contractually bound.”Slack of alternatives disappears
More detailsHide details

In a stable world

What I doI consolidate suppliers onto proven partners. A second supplier looks to me like a needless cost and overhead. Deeper integration with one of them reads as partnership. I do the same today with one AI model from one provider.
Why I do itI believe fewer and better suppliers mean lower cost and higher quality.
Why it made senseVolume lowers the price, shortens relationship management and raises quality, because the supplier learns your business. When you can switch if you need to, consolidation is clearly right.

In an exponential world

What happens under a shockFragility moves to the supplier, but you are the one paying for it. Once you stop investing in an area yourself, you lose the ability to judge what is worth anything in it — and you will not even notice that your supplier is falling behind (hard research: Cohen & Levinthal, Administrative Science Quarterly, 1990). Diversification is often only apparent as well: over 80 % of large European firms are at most three intermediaries away from a Chinese rare-earth producer (hard research: ECB Economic Bulletin 6/2025). The shock then arrives everywhere at once.
Counterfeit of resilienceAn SLA with penalties. „The supplier is contractually bound.“ A penalty will not deliver the goods, or the model that went down.
How to spot it in yourselfPick the supplier whose outage would hurt you most — including an AI model. How many weeks and how much money would switching to an alternative cost? If you do not know both numbers, the answer is „we don't know“. That is a result too.
New beliefI believe the cost of switching is part of the cost of supply.
New behaviourFor the most critical supplies I know the time and cost of switching, and from time to time I test a partial switch for real. With AI I own the way models are used and combined, not the model itself.
What it creates in the companyA known and paid-for way out. A stronger position towards a supplier who knows you can leave, and in-house expertise that notices when they fall behind.
Where to start on Monday: Write two lines: who your most critical supplier is, and who in your organisation knows how a replacement would be found. If the second line stays empty, you have your first task.

When it still holds: For commodities with a liquid market where switching is fast, and where depth of integration delivers quality two suppliers could not. Condition: you know the cost of switching and it is lower than the saving from consolidation.

Another seven

They did not fit on the stage. We meet them in companies just as often.

8
A question from the start of the talk

Exceptions through me

Anything non-standard is decided by the leader.

Counterfeit of resilience: “We have a clear escalation path and everyone has a deputy.”Decision slack disappears
More detailsHide details

In a stable world

What I doEverything that departs from the standard is decided by me. I delegate tasks, not decision rights. A question someone two levels down would answer in a minute sits in my inbox until the evening.
Why I do itI believe the best decision is made by the person with the broadest overview — and that is me.
Why it made senseThe leader has the most context. When conditions change slowly, their judgement is the best available, decisions stay consistent and the company is protected from expensive mistakes made by people who cannot see the whole.

In an exponential world

What happens under a shockContext changes faster than it can be carried upwards, so decisions are made on stale information. On top of that the leader is the queue for every exception, and their capacity is the ceiling on the speed of change. Frequent intervention from above pushes managers to escalate even more (consulting study: Sull, Homkes & Sull, Why Strategy Execution Unravels, HBR 2015). In a crisis everything is an exception, and the leader is the queue everyone is waiting in.
Counterfeit of resilienceAn escalation matrix and deputies. „We have a clear escalation path and everyone has a deputy.“ The bottleneck moved one floor. It did not disappear.
How to spot it in yourselfTake five decisions that came to you last week. In how many of them did somebody other than you know most about the matter? Additionally: how many hours on average did each wait for your decision.
New beliefI believe the best decision is made by the person closest to the problem — and my job is to give them the criteria.
New behaviourInstead of the task I delegate the criterion. For every type of recurring decision I write one sentence, „Decide it yourself if…“, with a clear boundary. Now and then I switch my phone off for a whole day and watch where decisions flowed.
What it creates in the companyDecisions where the information is, at a time when they still have value. A leader with capacity for the things nobody else can decide — and an organisation that works even when they are not available.
Where to start on Monday: Take the type of decision that reaches you most often and write one sentence for it: „Decide it yourself if…“. Send it to your team today.

When it still holds: For irreversible decisions where a mistake threatens the existence of the company or its licence. And where regulation assigns named accountability.

9

The line is certainty

Agreements across units go through the bosses.

Counterfeit of resilience: “Everyone knows what they are accountable for.”Decision slack disappears
More detailsHide details

In a stable world

What I doAnything that has to be agreed across units goes through the bosses. A direct agreement with a colleague from another unit reads to me as bypassing the line. Targets and bonuses are set vertically and cross-unit cooperation has no owner.
Why I do itI believe reliable coordination runs through the accountable line, because that is the only place it can be enforced.
Why it made senseThe line gives clear accountability and a traceable record. When the interfaces between units are stable and few, coordinating through a boss is cheaper than a web of ad-hoc agreements.

In an exponential world

What happens under a shockValue flows across units, but commitments hold only vertically. 84 % of managers can rely on their boss and their direct reports, while only 9 % can always rely on colleagues in other units (consulting study: Sull, Homkes & Sull, HBR 2015). People compensate by duplicating work and by hedging their promises. In a crisis that does not respect unit boundaries, the only route to coordination leads through two bosses — and both of them are sitting in the crisis team.
Counterfeit of resilienceAn org chart with clear responsibilities. „Everyone knows what they are accountable for.“ Vertically, yes. Horizontally, nobody.
How to spot it in yourselfTake the last three things that were delayed because of another unit. In how many of them did you talk directly to your counterpart before sending it upwards? For comparison: only 9 % of managers can always rely on colleagues in other units (consulting study: Sull, Homkes & Sull 2015).
New beliefI believe reliable coordination runs across — and my job is to make a horizontal commitment as enforceable as a vertical one.
New behaviourFor the most important flows of work that cross several units I name one owner of the whole flow and measure them by the result of the flow. Commitments between units have a name, a date and are visible.
What it creates in the companyCross-unit commitments that hold without two directors signing off. Work flows where value is created, and in a crisis there is more than one route to coordination.
Where to start on Monday: Call your counterpart in the unit you wait for most often and agree one specific thing directly — with a name and a date. Inform the bosses afterwards.

When it still holds: For allocating shared resources and where the law requires a named approval line. What should go through the line is the allocation of resources, not the agreement about the work.

10

Pressure as the lever

When results don't come, more pressure is applied.

Counterfeit of resilience: “We know how to dig in when it's needed.”Capacity slack disappears
More detailsHide details

In a stable world

What I doWhen the result does not come, I add pressure: a shorter deadline, more frequent checks, more hours. Stopping to find out why it is not working looks like a waste of time.
Why I do itI believe that when the result does not come, what is missing is effort.
Why it made senseWhen the process is capable and the slippage comes from insufficient effort, pressure raises performance quickly and reliably. The leader gets immediate proof that it worked.

In an exponential world

What happens under a shockThe extra effort consumes exactly the capacity from which understanding would have come. The short-term improvement masks a declining capability, so next time more pressure is needed for the same effect (hard research: Repenning & Sterman, 2001). Under pressure attention also narrows and people hold on to the last known procedure — exactly when a new one is needed [not verified for this material].
Counterfeit of resilienceCrisis mode. „We know how to dig in when it's needed.“ Every push is paid for out of the time for learning.
How to spot it in yourselfFor the last three slippages you handled by adding pressure: how many times did the team hold the new pace after the pressure was released, and how many times did it fall back? Write down both numbers.
New beliefI believe that when the result does not come, what is missing is understanding — and pressure is the most reliable way to prevent it.
New behaviourAfter a repeated failure I do not add pressure. I stop the team for a short reorientation: what we do not know about the problem and how we can find out cheaply.
What it creates in the companyUnderstanding of why things are not working, instead of a short-term push. Capacity that does not shrink after every crisis but grows.
Where to start on Monday: For a project that is slipping right now, do not add another checkpoint. Meet the team for half an hour with a single question: „What do we not know about this problem?“

When it still holds: When the cause of the slippage is known, the solution is known and only the execution is missing: a hard external deadline, a seasonal peak. The test: after the pressure is released, performance stays at the new level.

11

Everything is priority one

Whatever gets approved is started right away.

Counterfeit of resilience: “We have an overview of all projects.”Capacity slack disappears
More detailsHide details

In a stable world

What I doWhatever gets approved gets started. The number of running initiatives has no ceiling, because stopping an approved project looks like admitting a mistake.
Why I do itI believe an approved project should be started, and that running things in parallel shortens the overall time.
Why it made senseWhen projects do not share people and do not depend on each other, nothing waits to start and people are well utilised.

In an exponential world

What happens under a shockAverage lead time equals work in progress divided by throughput (hard research: Little's law; Little, Operations Research, 2011). Every additional project extends the time of all the others, and switching shared people lowers throughput itself. Nothing waits at the start, everything waits in the middle. When a shock arrives, stopping means twelve decisions with twelve sponsors.
Counterfeit of resilienceA portfolio review. „We have an overview of all projects.“ An overview is not a ceiling.
How to spot it in yourselfCount the initiatives that are started and unfinished in your unit today and compare that with the number of people working on them. Then ask three of those people how many different things they are working on this week.
New beliefI believe value is created only on completion, and that starting one more thing slows down everything already running.
New behaviourMy portfolio has a work-in-progress limit and a „one in, one out“ rule. I decide priority at the start of work, not at approval.
What it creates in the companyShorter delivery times across the board, not just for the priority. People who finish instead of switching — and free hands when a shock arrives.
Where to start on Monday: Pick one running initiative you would not start again today and put it on a list to be paused. If it is yours to stop, stop it.

When it still holds: When projects share neither people nor systems.

12

Calm is proof of health

A period without an incident is read as proof of safety.

Counterfeit of resilience: “We have a plan.”Learning slack disappears
More detailsHide details

In a stable world

What I doI read a period without incidents as proof that the system is fine, and as a reason to save on reserves and controls. A green dashboard means all is well.
Why I do itI believe that if nothing has happened, the system is fine.
Why it made senseIn a stable environment, time without failure really is a good estimate of reliability. All of reliability statistics rests on that. Reducing controls is then rational.

In an exponential world

What happens under a shockEvery calm day raises confidence at exactly the moment when hidden risk is growing. A turkey fed for a thousand days is at its most confident the day before Thanksgiving (opinion: Taleb, The Black Swan, 2007). The metric also replaces the goal: managers unknowingly substitute it for the strategy itself, and the more they are rewarded for it, the more they do so (hard research: Choi, Hecht & Tayler, The Accounting Review, 2012). Green then measures not the state of the system, but the time since the last shock.
Counterfeit of resilienceA business continuity plan. „We have a plan.“ A plan that has never been run for real is a hypothesis.
How to spot it in yourselfWhen did you last deliberately switch off something that works, to find out what happens? Write down the date. If the answer is „never“, your only test will run on a date you do not get to choose.
New beliefI believe a calm period is the least informative state I can be in.
New behaviourRegularly and in a controlled way I remove one established certainty and watch what fails: a day without a key person, a week with a backup supplier, an outage of a tool.
What it creates in the companyKnowledge of your own weak spots before the market finds them. A continuity plan that has been tested, not just written.
Where to start on Monday: Pick one system, supplier or person without whom „it wouldn't work“ and write down three things that would happen on day one of their outage. Tomorrow, show it to someone on the team and ask whether they agree.

When it still holds: Where the conditions have not changed over the observed period: physical wear, classic insurance risk.

13

Back to normal

After a crisis the original state is restored and a control is added.

Counterfeit of resilience: “We learned our lesson.”Learning slack disappears
More detailsHide details

In a stable world

What I doAfter a crisis I want the original state restored as fast as possible. The root cause analysis ends with a corrective action, usually one more control or approval step. What to change permanently is never asked, because operations are running again.
Why I do itI believe we handled the crisis once we got back to where we were as quickly as possible.
Why it made senseWhen the crisis was a deviation from a good state, a fast return is the cheapest option. A control added at the point of failure is cheap and effective if the same cause can come back.

In an exponential world

What happens under a shockA crisis is the most expensive experiment a company will ever get for free, and going back to normal throws it away. Every crisis adds a control, so the system gets heavier cycle by cycle. After a real shock the original state often no longer exists, and the return is the restoration of something that has ceased to hold. Ecology distinguishes stability, the ability to return to equilibrium, from resilience, the ability to persist even through a change of state — and warns that these are not the same property (hard research: Holling, Annual Review of Ecology and Systematics, 1973; open-access PDF, IIASA).
Counterfeit of resilienceA post-mortem with a list of corrective actions. „We learned our lesson.“ The lesson is a new control, not a new system.
How to spot it in yourselfTake the last three crises or serious incidents. How many rules, controls or approval steps did you add after them, and how many did you remove? Write down both numbers.
New beliefI believe we handled the crisis once we permanently changed at least one assumption we had been standing on.
New behaviourEvery post-crisis review has a mandatory question: what are we changing permanently and what are we removing. For every new control step I look for one to remove.
What it creates in the companyA system that moves forward after every crisis instead of getting heavier. The crisis stops being a pure loss and starts paying itself back.
Where to start on Monday: Open the review of your last crisis. For each corrective action ask: „Did this change an assumption we were standing on, or did it only add a control?“ Propose one control-only action for removal.

When it still holds: When the cause was external and one-off and the system is demonstrably sound. And in the acute phase, when getting back into operation simply is the priority.

14

Training will fix it

When people can't do something, they are sent to training; the environment stays.

Counterfeit of resilience: “We invest in our people.”Learning slack disappears
More detailsHide details

In a stable world

What I doWhen people cannot handle new things, I send them to training. The system around them — targets, processes, time, what I praise them for — stays the same.
Why I do itI believe that when people cannot do something, it is enough to teach them.
Why it made senseWhen the work changes slowly and the environment supports the new behaviour, training is the fastest route to a skill. What was learned reinforces itself on the job.

In an exponential world

What happens under a shockWhether what was learned transfers into the work depends heavily on the work environment — support from managers and peers and the opportunity to use the new things (hard research: Blume, Ford, Baldwin & Huang, Journal of Management, 2010, meta-analysis). When the environment does not change, the behaviour goes back. And meanwhile the exponential world changes what needs to be known faster than the next training can be planned. The often-quoted figure that only 10–15 % of training transfers into practice could not be verified in the source and its origin is disputed. Not used here.
Counterfeit of resilienceA development plan and the number of training hours delivered. „We invest in our people.“ Hours in a classroom are not a change of behaviour.
How to spot it in yourselfTake the last three trainings your team attended. In how many of them did something other than people's knowledge change in the work right afterwards — time, a rule or a measure?
New beliefI believe people do what the environment allows and rewards. Training without a change of environment is an expense, not an investment.
New behaviourFor every training I name one thing in the work that changes right afterwards: time to practise, a rule, a measure. If I cannot name it, I do not order the training.
What it creates in the companyNew behaviour that lasts, because the environment supports it. Development money that shows up in the work, not only in the count of training hours.
Where to start on Monday: For your next planned training, write down one thing in the work that will change a week after it. Send it to the participants before the training.

When it still holds: For knowledge with a clear right answer and immediate use: safety, regulation, working with a specific tool.

We are happy to send you the full catalogue with examples from banking, IT and manufacturing and with all the sources.

Send me the full version
White segment: 30 % licences and infrastructure

Measuring the business value of AI adoption

AI transformations swallow a lot of money, but the impact on profitability often cannot be proven. Licences and infrastructure make up only 30 % of the budget. The remaining 70 % is process change and adoption by people, and that often is not measured at all.

The article lays out the Input, Output, Outcome, Impact framework that separates a provable benefit from AI overhead. It adds ten hidden costs that nobody sees at planning time, and a procedure for companies that already have AI and have not measured it so far.

Read the article

Jan Krchňák, RainFellows

Interested in the topic?

Would you like to know more, or talk about the situation in your own company? Let us discuss it over coffee, in person or online. Leave us your contact details and a few words about what you are dealing with.

Tomáš Tureček or Martin Chmelař will get back to you.

Contact us