Crafting Conflict In An Illegible Org

10 min read Original article ↗

Leadership under load is not about being inspirational. It is the work of steering systems you cannot fully see, with people who are already tired, under incentives you did not design.

This post sits in the Leadership Under Load series and ties Crafting Conflict Vol 1 to The Illegibility Crisis. Leadership Under Load is the ground where systems become illegible and conflict becomes the last instrument you can trust.

That is the overlap between the two books. Crafting Conflict Vol 1 sits in the room with you when tension shows up and starts to bend the work. The Illegibility Crisis(currently a work in progress) sits over your shoulder when AI, vendors, and internal tools blur your view of the systems themselves. Same problem, different zoom.

Conflict is the first honest signal you get that your instruments are lying, or at least missing something important.

You can ignore that signal, punish it, or use it. I wrote both books to make "use it" less mystical and more repeatable.


When Conflict Is The Only Telemetry You Can Trust

You see versions of this in hospitals, banks, public agencies, anywhere AI and complexity outpace legibility. In an AI heavy org, the story often looks like this: Output looks fine. Dashboards green. Docs are clean and written in familiar templates. Promotion packets look impressive.

Then a real incident hits or a crown jewel project stalls and you find out the people who shipped the senior looking work cannot reason about it under pressure. No one can cleanly trace who actually decided what. The last person who really understood the old system is two reorgs away.

That is the illegibility crisis: output has stopped reliably telling you what it used to mean about understanding, judgment, or risk. When that happens, conflict is often your loudest, clearest signal that something structural is off. A senior IC and a manager are locked in a loop about "velocity vs safety". Platform, data, and security keep having the same quiet proxy war about "blocking" vs "protecting".

Reviewers are quietly avoiding certain repos because they do not trust their own mental model anymore.

Those fights are not about personality. They are about fractures in how knowledge, power, and risk are lined up. If you treat them as noise, you lose the only telemetry you still have.

Two Books, One Job

The books are aimed at the same work from different sides. Crafting Conflict names the six saboteur patterns that show up under load: Cynic, Ghost, Hero-Martyr, Passive Burnout, Underminer, False Compliant. It treats those patterns as system responses to repeated harm and drift, not as permanent character defects. It gives you field tools for seeing the pattern, containing the damage, and giving people a real way back into the work.

The Illegibility Crisis names the fractures that make your org hard to read: Synthetic Competence, Decision Fog, Knowledge Drift, Ghost Apprenticeship, Promotion Blindness. It forces you to draw a hard circle around a few crown jewel systems where you refuse to be ignorant. It gives you four verbs instead of slogans: map it, probe it, trace it, teach it.

Together, they describe one discipline. You cannot lead what you cannot see. You cannot see what you are not allowed to talk about. You cannot talk cleanly about anything if you have no craft for handling conflict.

Conflict As A Crown Jewel Instrument

So how do you actually use conflict as an instrument instead of treating it as noise? The wrong move here is to add more dashboards and zero new conversations. In a relatively legible org, conflict is cheap. You can assume most people looking at the same system are at least seeing the same thing. The argument is over priority or taste.

In an illegible org, conflict is expensive and diagnostic. Every spike of tension around a critical system is telling you at least one of three things: we do not share a map, we do not share a story of how decisions get made, or we do not agree who is carrying which risk.

You can ignore conflict around low stakes work. You cannot ignore it around crown jewels without betting the company. If the system failing would force you to explain it to a board or a regulator, it is a crown jewel.

So the move is not "smooth it over and ship the next sprint". The move is to name the system and the fracture. Which crown jewel are we actually fighting about. Which fracture is showing up: Synthetic Competence, Decision Fog, Knowledge Drift, Ghost Apprenticeship, Promotion Blindness. If you cannot answer that, you are arguing about mood, not risk.

Then name the saboteur patterns you are seeing. From the Crafting Conflict side, that means looking at the behavior, not the story people tell about themselves. Is this Cynic energy, spotting real failure but stuck in "nothing will change". Is this Ghost energy, staying quiet to stay safe and starving you of signal. Is this Hero-Martyr, over functioning and centralizing everything "because no one else can". Is this Passive Burnout, present but checked out, quiet erosion of care. Is this Underminer, agreeing in the room and then attacking the decision through sarcasm and side sniping. Is this False Compliant, mirroring all the right language, then reverting once the heat is off.

In practice you get combos. Cynic + Ghost in the same room feels like sharp private analysis and public silence. Underminer + False Compliant feels like conflict theater that never actually resolves. The goal is not to label people. The goal is to put words on the pattern so you can treat it at the right scale.

Then turn what you just learned into instrumentation. From the Illegibility side, that means taking the conflict seriously as data. Map it: what system, decisions, tools, vendors, and roles are actually in play. Probe it: what question did this conflict finally force into the open. Trace it: whose judgment is actually shaping the current behavior, on paper and in practice. Teach it: what you want the next person in this role to know before they inherit this exact fight.

Conflict stops being an HR problem and becomes one of your core instruments for reading an org your dashboards no longer describe.

What This Looks Like In Practice

Three of the most common feedback loops I see under load.

Promotion under scaffolding. The argument is "Are they actually senior, or just good at shipping with AI scaffolding." In the room you usually have Cynic ("we are promoting PowerPoint, not judgment") and Ghost (people who disagree but stay quiet because they do not want to be the problem). In the system you are in Promotion Blindness and Synthetic Competence territory.

The Crafting Conflict move is to say out loud that the real question is judgment, not volume. Catch the Underminer pattern early if someone starts sniping about "favorites" instead of naming criteria. The Illegibility move is to build a small set of judgment probes and crown jewel scenarios into the promotion process. Make the decision record explicit: which calls this person made, on which systems, under which constraints. You use the conflict as design input for the instrument.

Every time you promote the engineer who shipped 50 AI-assisted PRs but cannot explain why the system behaves differently under load, you reinforce Synthetic Competence and make the next promotion fight harder to read. Over time that loop trains your org to favor Synthetic Competence unless you break it on purpose.

Incident blame loops. The argument is "Was this a bad process, a bad call, or a bad actor." In the room you might see Hero-Martyr ("I will just carry more, next time I will catch it") and Passive Burnout ("this is just how it is now"). In the system, you are staring at Decision Fog and Knowledge Drift.

The Crafting Conflict move is to ban character assassination in the review. Separate "what hurt" from "what happened" and "what needs to change". The Illegibility move is to use the conflict to trace where real authority actually lived at each step. Map which parts of the incident no one can explain without pulling logs or AI generated artifacts, then treat that as a risk, not a curiosity.

Every incident you close without tracing real authority deepens Decision Fog and makes the next crisis harder to resolve. Again, conflict is the entry point, not the distraction.

Vendor wars. The argument is "This vendor is killing us" vs "We cannot replace them right now". Common pattern: Cynic in engineering, Ghost in leadership, Underminer in the middle. System fracture: Knowledge Drift into the vendor console, plus Ghost Apprenticeship where the real skills to operate core systems live with a few specialists.

The Crafting Conflict move is to surface the fear explicitly instead of sniping. Call out False Compliant behavior where people say "we are fine" to avoid another hard project. The Illegibility move is to treat the vendor surface as part of a crown jewel, not a black box. Map which capabilities are effectively outsourced and what that does to your incident posture and promotion ladder.

Every quarter you defer this mapping, the Knowledge Drift compounds and the exit cost of the vendor rises. These are not isolated events; they are feedback loops that train your system to prefer illegibility over honest assessment.

The Price Of Using Conflict This Way

Using conflict as instrumentation has a cost. You will surface tensions everyone has quietly agreed to work around. You will find out that some "top talent" is mostly synthetic output sitting on shallow understanding. You will see which systems are effectively black boxes no one can explain until something breaks. When one of those systems fails at 3am with a patient in the room or money in flight, you will find out exactly what illegibility costs.

Using conflict as instrumentation costs attention and political capital you might prefer to spend elsewhere.

It is tempting to keep the appearance of peace and keep your illusions. That temptation is not neutral. It is a choice about what kinds of incidents and exits you are willing to accept later. Leaders are not exempt; an executive team can run False Compliant too, praising "candor" while quietly rewarding only smooth status.

The cost of not doing this is quieter, until it is not. You keep promoting based on artifacts instead of judgment. You keep shipping AI-assisted changes into systems no one can trace. You keep running incidents where the real decisions live in tools and never make it into words.

Conflict is the instrument that refuses to let you lie to yourself about any of that.

A Simple Field Test For This Week

The boon is simple: one conflict you choose and contain, instead of ten the system chooses for you. Pick one live conflict in your world. Not the biggest one. One that keeps showing up in status.

Run this three step pass:

  1. Name the fracture. Name the system gap, not the person. Which Illegibility fracture does it smell like.
  2. Name the saboteur pattern. Is this Cynic, Ghost, Hero-Martyr, Passive Burnout, Underminer, False Compliant, or a mix. What is the cost if it keeps running for six more months.
  3. Design the smallest conflict play. One meeting, one decision, one experiment. Timebox it. Write down what "better" and "worse" look like by a specific date.

Then ask one blunt question:

If we pour AI on top of this exact fracture, does it get better or worse.

If the answer is "worse", that is your next leadership move. Not another AI mandate. Not another pep talk. One crafted conflict, on purpose, before the system does it for you.