Twelve Conditions for the Crown Fire

An executive primer on reading network readiness — three of twelve dimensions in the wildfire register that the rest of the framework inherits.

Wildland fire crews spend a lot of time not fighting fires. They walk the back country before fire season, reading conditions: how much fuel is on the ground, how dry it is, where the ladder fuels are dense enough that surface fire could climb into the canopy, where the wind funnels through ridge gaps, where the last fire ten or twenty years ago left a mosaic of older-growth and younger regrowth. The walk produces a fuel-readiness map. The map does not predict whether the forest will burn. It predicts where the fire will jump from surface to crown when it comes, because some fire is always coming.

I have started thinking about the twelve-factor AI-readiness self-assessment that anchors part of my consulting practice the same way. The dimensions — management buy-in, change fatigue, workflow mapping, organizational design, trust in AI, transparency, technology capability, restructuring, talent-strategy posture, partnership framing, governance, QA discipline — are not pass/fail conditions for whether to deploy AI inside an organization. They are fuel-readiness readings for where the rollout will catastrophically fail when it inevitably encounters the early majority of the organization.

The conventional move in the consulting practice is to score the twelve, find the lowest three or four, and prescribe interventions that bring those scores up before deploying.¹ The conventional move treats the instrument as a gating function — you cannot ship until the scores look like the scores of the firms whose AI rollouts have succeeded. The conventional move is wrong, in the specific way the 10 a.m. policy was wrong about wildfire suppression for fifty years: it treats a landscape variable as a local emergency. The 12-factor reading is a fuel map. The right question to ask about it is not what is our lowest score? The right question is where does our fuel ladder break, and which dimensions tell us where to set the seed cohort?

I am going to walk three of the twelve dimensions in their reframed form. The pattern across the other nine works the same way. The point of walking three is not to teach you all twelve; it is to teach you the move — the shift from a population-level signal to a network-position signal — so that when you read the remaining nine you do not need anyone to explain them.

What the twelve are doing, mechanically

The original instrument scores each dimension one through five at the organizational level. How is our management buy-in? Three. How is our change fatigue? Four — meaning bad. How is our workflow mapping? Two. The output is a radar chart that consulting practices have been monetizing for two decades, sometimes with twelve dimensions and sometimes with eight or sixteen, but always with the same fundamental shape.

The implicit theory behind the radar chart is that organizations have aggregate-level properties that determine adoption outcomes. Treat the organization as a single object; score the object on each dimension; act on the dimensions where the score is lowest. The theory is wrong in roughly the way the firm-as-a-black-box models from mid-century industrial economics were wrong — the firm is not a single object; the dimensions are not aggregate-level properties; the dynamics that determine adoption operate at a smaller unit than the firm.

That smaller unit is the tie cluster — the local group of people whose adoption decisions reinforce each other because their work and their trust ties are dense enough that what one does, the others see and respond to within the same week. A 50,000-person organization has somewhere between two and five thousand of these clusters depending on the work structure. Adoption happens or fails to happen inside the clusters, not at the level of the firm, because the contagion that matters in adoption is complex — it requires multiple reinforcing exposures from trusted ties.² An aggregate-level score that averages across the clusters misses the variance that determines whether adoption actually propagates. The score can be three out of five on management buy-in, and you can have one cluster where every manager is a structural bridge with strong ties to the seed cohort and another cluster where every manager is isolated from it; the two clusters will produce opposite outcomes from the same score.

The reframe of any one of the twelve dimensions is the same move applied locally. Where in the network is this dimension high? Where is it low? What does the spatial distribution tell us about where the rollout will catch and where it will spot-fire into bare ground? That is the move. Three demonstrations follow.

One — Management Buy-In

The conventional treatment is unambiguous. You survey the management layer; you ask three or four questions about enthusiasm for the AI initiative, perceived value, willingness to advocate; you compute a score; you intervene on the managers who score lowest. Targeted retraining, one-on-one executive coaching, sometimes performance-management consequences. The implicit theory is that management buy-in propagates downward — once the manager is bought in, the team follows. The CHRO who has run this play more than once knows what comes next: the manager who scored high in the survey sometimes produces a team that does not adopt; the manager who scored low sometimes produces a team that does. The score has predictive value only at very wide aggregates. Inside the variance, it is almost noise.

The reframe is unflattering to the survey instrument: management buy-in is the wrong unit of analysis. What predicts whether a manager's team adopts is not the manager's enthusiasm at the time of survey. It is the manager's network centrality — how many of the manager's peers and the manager's team members are themselves in the seed cohort or in adjacent already-adopted clusters — and the tie strength between the manager and that already-adopted network. A manager with high enthusiasm and low centrality is a buy-in mirage. A manager with moderate enthusiasm and a dozen strong-tie bridges into the early-adopter cohort is the actual lever; that manager's team will adopt whether or not the manager personally is a champion, because the team's trust ties give them the multiple reinforcing exposures the adoption requires.

What you instrument, instead of the survey: an organizational network analysis pass — the kind Chapter 6 of Backburn spends a chapter on — that identifies which managers sit in structural-bridge positions relative to the seed cohort.³ Score each manager not on their stated enthusiasm but on their tie distance to the seed cohort, weighted by tie strength. The map this produces does not look like an HR scorecard. It looks like a fuel-continuity map: where managers form connected bridges to the seed cohort, the rollout will propagate; where managers sit as structural isolates, the rollout will stall regardless of how much you retrain them.

The operational consequence is the move every consulting playbook fails to make: stop trying to convert the structural isolates. Their buy-in is not the lever. Instead, reinforce the bridges that already exist, by giving the bridge-positioned managers the time and the social permission to participate in the seed cohort's working sessions, even if their formal sponsorship of the initiative is moderate at best. The bridge will carry the adoption across distances the isolated champion never could. This is not a soft observation. The buy-in score and the network-readiness score are measuring different objects — one of them is calibrated to stated enthusiasm, the other to structural position — and only the structural-position instrument has a defensible mechanism for predicting what happens after rollout. The empirical replication record on attitudinal-readiness scores in adoption settings, surveyed in the diffusion-of-innovations literature since Rogers's third edition, is consistently weak; the empirical record on network-position measures, in Burt's program and Centola's program both, is consistently strong.

Two — Change Fatigue

The conventional treatment is a pacing decision. You measure organizational change fatigue with a sentiment survey, you weight it against the perceived urgency of the new initiative, and you sequence rollouts to avoid concentrating change in fatigued populations. We will hold the AI launch until next quarter because the workforce just absorbed the new HRIS migration. The implicit theory is that fatigue accumulates uniformly across the organization — every person carries a roughly equal load of attention to change, and once the aggregate exceeds a threshold, the organization is change-saturated and further rollouts will fail.

The reframe is the structural-bridge problem from a different angle. Fatigue does not accumulate uniformly because change does not arrive uniformly. In multi-initiative organizations more broadly, a small fraction of the workforce — the people in network-bridge positions — absorbs disproportionate fatigue because they are touched by every rollout. They are the cross-functional connectors; they sit in cross-team coordination meetings; they are the early adopters of the previous five tools because they were the people whose roles required them to integrate the previous five tools into the work of teams that did not. By the time the AI rollout begins, the bridge actors have been in adoption-mode for two or three years, and the rest of the organization has been in mostly-stable mode. The aggregate score reads organization is moderately fatigued. The reality is that ten percent of the organization is profoundly fatigued and ninety percent is not.

This matters operationally because the bridge actors are precisely the people whose adoption propagates the new tool. The conventional move pacing rollouts off the aggregate fatigue score is functionally indistinguishable from a rollout pacing that ignores fatigue entirely — the aggregate hides where the fatigue actually lives. What you instrument instead: bridge-load — for each individual, the count of distinct change initiatives that have touched their work in the previous twenty-four months, weighted by the duration each initiative required active attention. Combine with the tie-overlap map from the structural-hole pass. The output is a fatigue topology rather than a fatigue score, and the topology tells you which bridge positions need active relief before they can absorb the new rollout, and which can take more load without breaking.

The discipline the conventional playbook gets right, sometimes accidentally: a fatigued bridge is a worse adoption seeding position than a fresh isolate. If your seed-cohort selection from Chapter 7 has produced a roster that overlaps significantly with the bridge actors who carried the last three rollouts, the rollout will fail not because the cohort is the wrong selection — it is the right selection on every other criterion — but because the cohort is being asked to do another adoption-propagation job on top of the residual load from the previous three. The fatigue topology is what tells you to pull two or three of those cohort members for the new rollout's seeding job and pull the rest into a sustainment role for one of the older rollouts that is still in mop-up.

Three — Workflow Mapping

The conventional treatment is the most procedurally legitimate of the three and the most operationally misleading. You commission a workflow-mapping pass: process documentation, formal handoffs, system-of-record dependencies, the org-chart-shaped picture of how work moves through the organization. You identify the steps where AI insertion is technically feasible. You stage the rollout against those steps. The implicit theory is that the formal workflow is the real workflow, and AI inserted at any step in the formal workflow will encounter the same friction profile that human work at that step encounters.

The reframe is the one every operator who has actually watched a workflow happen knows in their gut. The formal workflow is not the real workflow. The real workflow consists of the informal coordination paths — who asks whom for help, which two people on different teams have a side-conversation about how to handle the edge case that the documented process does not cover, where the expediter role is in each team (the person who unblocks everyone else's stuck work, even when their role description says nothing of the kind), where the translator role is between technical and operational language. These informal nodes are where the work actually moves, and they are where AI insertion either lands as augmentation or fails as imposition.

Operationally this means workflow mapping done only against the formal process produces an AI insertion plan that targets the documented steps and ignores the nodes that actually carry the coordination. The cost: AI shipped against the documented steps interacts with humans who were not the people performing those steps, who treat the AI as an outside imposition because their unspoken contribution to the workflow has been ignored. The rollout encounters resistance that the formal mapping did not predict, the resistance gets read as change fatigue (see above, where it is misread the same way), and the AI insertion produces a worse coordination outcome than the unaided informal workflow.

What you instrument instead: an informal-workflow ONA pass, layered against the formal one, mapping the who-asks-whom-for-help network for each work domain the AI insertion will touch.⁴ The output is a different kind of map than the formal one — it has people in it whose formal role descriptions say nothing about coordination, and it has formal coordinators in it whose informal role is mostly performance of the formal role. You build the AI insertion plan against the informal map. The insertion lands at the nodes where the practical work is happening, and the practical work gets augmented. The formal map becomes the compliance overlay, not the design substrate.

What the three reframes share

The pattern across the three dimensions, and across the remaining nine in the instrument, is the same move repeated. Every dimension that the conventional consulting practice treats as an organizational property — management buy-in, change fatigue, workflow legibility, transparency, trust, restructuring readiness — is more usefully treated as a spatial distribution across the network. The aggregate score is almost always less informative than the topology that produces it. The conventional intervention — raise the low scores — is almost always less effective than the topological intervention — find the bridges and reinforce them; find the isolates and route around them; find the bridge-load concentrations and relieve them before adding more.

This is what the wildfire vocabulary the book is built on is for. Fuel load is not a single number for a forest; it is a continuity map across the forest. Fire line is not a single barrier; it is a sequence of anchor points and dozer cuts and natural features that together hold or do not hold. Backburn is not a single decision; it is a coordinated maneuver that depends on conditions across the entire fire's neighborhood. The crew that does this work has practitioner discipline that the conventional adoption playbook lacks, and the discipline shows up in the questions they ask: not is the forest ready to burn? but where will the fire jump from surface to crown when it comes? Not is the organization ready to adopt? but where will the adoption stall when it encounters the early majority — and which dimensions of fuel-readiness tell us where to set the seed cohort to keep that stall from becoming a crown-fire failure?

The twelve-factor instrument, read in this register, becomes a fuel-readiness map for the rollout's fire season. That is what it should have been all along.

The practitioner's question

When a fire behavior analyst stands on the line and reads the conditions, they are not asking whether the forest deserves to burn. They are asking where the fuel ladder breaks. The break is where the surface fire's energy runs out before it can climb to the canopy; it is the structural feature that turns a survivable fire into a survived one.

I am asking the equivalent question of every executive I work with: where does your network-readiness ladder break, and what does that tell you about where to seed the cohort that will carry the AI rollout through the chasm? It is a different question than is your organization ready? It is more specific. It is more useful. It produces an action.

There are nine more dimensions in the instrument, and walking them in this register would produce the same shape of analysis nine more times. I am not going to do that here. I am going to trust that any executive who has read this far can do the move for themselves on the remaining dimensions — and trust that the practitioner who can do the move is the practitioner the AI era will reward.

A fire crew, walking out at the end of the day before the burn arrives, knows which ridges they will hold and which they will give up to the fire. The crew has the discipline because they have the map. The crew has the map because they took the walk. The discipline lives in the walk.

Take the walk.

The self-assessment instrument was designed for AI-readiness consulting; the dimensions were derived from the organizational-adoption literature and from the convergent failure-mode framings documented in the consulting white-paper corpus catalogued at content/research/ai-encyclopedia/manifests/white-papers-2026-05-12.json. The instrument's full dimension definitions live in content/research/ai-encyclopedia/manifests/self-assessments-2026-05-12.json. The reframe-as-network-readiness move documented here is from 2026. ↩
Damon Centola, How Behavior Spreads: The Science of Complex Contagions (Princeton University Press, 2018), chapters 2 and 4. The complex-contagion distinction underwrites every reframe in this piece; if you read one of the citations in this footnote system, read this one. The 2010 Science lab experiments and the 2018 synthesis have since been independently replicated at country scale: a 2025 field RCT in Sociological Science (volume 12) used a peer-encouragement design to test the multi-source-reinforcement claim at population scale and found strong support — signals from multiple ties interact rather than acting as independent cascades. The lab-to-field gap on the central mechanism is closed. ↩
Ronald S. Burt, Structural Holes: The Social Structure of Competition (Harvard University Press, 1992). The structural-hole concept is the academic anchor for the bridge-vs-isolate distinction this section turns on; Burt's later Brokerage and Closure (Oxford, 2005) extends the argument to the value a bridge captures by sitting between disconnected clusters. ↩
The relevant academic anchor here is Rob Cross's work on organizational network analysis as a coordination instrument — particularly The Hidden Power of Social Networks (Harvard Business School Press, 2004, with Andrew Parker). The methodology is well-developed; what is missing in most consulting practice is the discipline to use it before the rollout rather than as a post-mortem on a rollout that failed. ↩