Glossary

Glossary

The book's operational vocabulary at your fingertips.

Agentic / Autonomous agent: An AI system able to carry out complex, multi-step tasks with a high degree of autonomy in planning and decision-making, as opposed to a copilot that simply assists a human on a single task. Its spread since 2024–2025 is reshaping the nature of cognitive work.
See chapter 1
Agility Committee: The governance body of the system, bringing together at least four voices: business, risk or compliance, HR, and legal. In zones of high collective impact, an employee representative joins as well. The committee arbitrates on the decisions the technical system proposes — without settling them alone.
See chapter 8
Allostatic load: A concept from occupational medicine formulated by Bruce McEwen at Rockefeller University in 1998. It refers to the cumulative metabolic wear caused by continuous exposure to changes in context without periods of stabilization. Invoked in the book to justify stabilization plateaus.
See chapter 15
Arbitration visa: The second of the two cumulative visas required to trigger Pay-for-Agility. After-the-fact validation, by the Agility Committee, that the rewarded contribution is documented, measurable, and reversible. The absence of either visa blocks the bonus.
See chapter 12
Asymmetry of application by zone: The fourth pillar of the protection doctrine. The book's frameworks do not apply uniformly across all zones, but with differentiated intensity: full relevance in Laboratory and Front Line, withdrawal in Fortress. This asymmetry is what shields the doctrine from uniform application.
See chapter 8
ATS: Applicant Tracking System. Software for managing applications in external recruitment. Traditional ATSs filter résumés on keywords and formal criteria, producing a paper-ceiling effect on atypical profiles whose trajectory the grids cannot read.
See chapter 4
Bastion: One component of the leader's personal support architecture, structured in four parts: an operational steering trio, an institutional challenger, a technology-ethics arbiter, and a cognitive decompression chamber. Together, the four bastions replace the myth of heroic solitude with an auditable discipline of resilience.
See chapter 15
Capacity to unlearn: The fourth layer of the Potential Stack. An employee's ability to let go of a mental model that was right for a long time, in order to adopt another that the new environment makes more relevant. Observable in transitions between contexts and in shifts of technological paradigm.
See chapter 6
Cognitive decompression chamber: The fourth bastion of the leader's personal support architecture. A strictly personal, non-institutional space: peer supervision from other industries, senior mentoring, dedicated professional support. It lets the leader process the anxiety of continuous change without projecting it onto the workforce.
See chapter 15
Cognitive flexibility: The first layer of the Potential Stack. An employee's ability to hold several reasoning frames at once, to shift perspective depending on the problem, and to transfer a line of reasoning across domains. This is not the same as versatility — it names a quality of the mental activity itself.
See chapter 6
Contractual imperviousness: The first pillar of the protection doctrine. The dynamic reading system alters neither the employment contract, nor the collective pay grid, nor the classification, nor any acquired rights. The opportunities the radar surfaces never trigger a sanction. That imperviousness is the condition of legal and social legitimacy.
See chapter 8
Crash-test: A methodical exercise testing the model against its nine main objections — technical, legal, social, and ethical. Run according to a protocol that separates what the model concedes, what it contests, and what stays fragile despite the answer. It precedes any rollout.
See chapter 13
DPO: Data Protection Officer. A role made mandatory by the European GDPR and by several converging frameworks. In the book's doctrine, the DPO acts as the technology-ethics arbiter and holds a veto over any use drifting toward intrusion or indirect discrimination.
See chapter 8
EU AI Act: The European regulation on artificial intelligence (2024/1689), whose full obligations for high-risk systems became enforceable on 2 August 2026. AI systems used for recruitment, selection, or performance evaluation fall into this category and require human oversight, transparency, and appeal procedures.
See chapter 8
Expert ceiling: A situation in which the depth that once made an expert valuable becomes insufficient for an environment whose structure has changed — with neither the person nor the organization able to convert that expertise. Not a fault, not an age effect: the result of a trajectory meeting a shift in context.
See chapter 3
Exploratory curiosity: The third layer of the Potential Stack. An employee's inclination to explore domains adjacent to their original job — to read, learn, and experiment beyond their strict scope. Judged by what the employee does with their learning, not by the volume they consume.
See chapter 6
F/FL/L doctrine: The common shorthand for the Fortress / Front Line / Laboratory triad. The risk-zone differentiation doctrine that grounds the asymmetric application of the book's frameworks.
See chapter 8
Fast learner: A profile type identified by the Trajectory Radar and the Reverse People Review. An employee whose measured time to adoption is 50% shorter than their department's average. Typically combines the first two dominant layers of the Potential Stack. Represents between 3% and 9% of a typical workforce, according to the deployments documented in the book.
See chapter 14
Fortress: The first of the model's three zones. Activities where the cost of error is high, the environment is relatively stable, and human responsibility is irreducible: industrial safety, heavy compliance, critical maintenance, surgery, binding regulatory validation. Stable depth and procedural vigilance are what prevail here. Conventional tools keep their relevance.
See chapter 8
Four pillars of protection: The system's defensive doctrine, built on four principles: contractual imperviousness, systematic human oversight, parameter transparency, and asymmetric application by zone. The simultaneous activation of all four is what conditions the legal, social, and operational legitimacy of any rollout.
See chapter 13
Framing visa: The first of the two cumulative visas required to trigger Pay-for-Agility. Up-front validation that the activity belongs to a Front Line or Laboratory zone, that it is part of a transformation project running at critical speed, and that the budget envelope is calibrated and bounded.
See chapter 12
Front Line: The second of the model's three zones. Activities where the cost of error is high and the environment evolves quickly: cybersecurity crisis management, mergers and acquisitions, compliance against shifting regulations, AI-augmented risk analysis. They call for a pairing of an anchoring expert and a fast-reconfiguring profile.
See chapter 8
HRBP: HR Business Partner. A field HR role attached to a business unit, tasked with translating central HR frameworks into the operational context of that unit. In the book's logic, HRBPs typically sit in a Front Line zone within the HR function itself.
See chapter 10
HRIS: Human Resources Information System. The layered architecture described in chapter 8 sets out an HRIS that separates the contractual layer (payroll, grid, collective agreements — all immovable) from the dynamic reading layer (opportunities, trajectories, governed by the Agility Committee).
See chapter 8
Hybrid team composition: A team built from three or four complementary profiles on the same critical project: a Fortress expert for regulatory depth, a fast learner for adoption speed, a knowledge transmitter for diffusion, and an unforeseen-events leader to arbitrate the tensions. It optimizes safety, pace, and transmission all at once.
See chapter 14
Knowledge transmitter: An employee who plays a structuring role in spreading knowledge within a team, turning individual mastery into a collective asset. They write the playbooks, stabilize the templates, and keep knowledge circulating. Their presence in a team roughly halves the collective Time-to-Skill.
See chapter 14
Laboratory: The third of the model's three zones. Activities where the cost of error is bounded, the environment evolves quickly, and experimentation is the natural path to value creation: R&D, new-product design, technology integration, emerging-market exploration. Learning speed and exploratory curiosity are what dominate here.
See chapter 8
Layered architecture: An HRIS organized in three distinct layers: the contractual data layer (System of Record), the intermediate aggregation and filtering layer, and the human decision layer. The separation is technical — but also legal and ethical.
See chapter 8
Learning agility: The second layer of the Potential Stack. An employee's ability to acquire new know-how quickly, to turn brief exposure into productive autonomy, and to transfer an acquired logic from one domain to another. Measured through Time-to-Skill.
See chapter 6
LXP: Learning Experience Platform. A training platform that favors self-directed learning and personalized paths, as opposed to traditional catalogs. The usage traces an LXP produces feed one of the observable signals of the Potential Stack and Time-to-Skill.
See chapter 9
Nine-box matrix: A nine-cell matrix crossing an axis of past performance with an axis of perceived potential, used since the 1970s to structure talent reviews. It remains acceptable in stable environments — in the Fortress zone — and becomes ill-suited as soon as reconfiguration runs fast.
See chapter 5
ONA: Organizational Network Analysis. A method that maps the real exchanges between employees — informal requests, documentation flows, ad-hoc collaborations — in order to identify the structuring roles the org chart misses. Used with care, it surfaces knowledge transmitters that classic systems leave invisible.
See chapter 8
Paper ceiling: The structural effect by which the diploma, the academic pedigree, or the formal career path stops being one signal among others and becomes an implicit ceiling on the visibility and recognition of trajectories. It operates in external recruitment and in internal promotion alike, producing a self-fulfilling prophecy.
See chapter 4
Parameter transparency: The third pillar of the protection doctrine. The weighting criteria of the algorithmic system are published, explained, defensible before employee representative bodies, and amendable through collective bargaining. That transparency is the condition for both internal and external audit, and for compliance with the 2026 regulations.
See chapter 13
Pay-for-Agility: A variable, targeted, reversible pay mechanism that rewards contributions to adoption speed in zones where that speed creates measurable value. It does not touch base salary, does not apply to Fortress zones, and is triggered only on contributions validated by the Agility Committee.
See chapter 12
PBO: Potential-Based Organization. An organizational architecture that adds to the SBO a layer for reading adaptation trajectories. The PBO does not replace the SBO — it sits on top of it. It brings in temporality, adoption speed, and transfer capacity, and belongs primarily to the Front Line and Laboratory zones.
See chapter 6
People review: The annual ritual by which leadership collectively reviews the placement of employees across the talent tools: evaluation matrices, high-potential pools, succession plans. The book proposes shifting that ritual, in part, toward the Reverse People Review.
See chapter 5
Potential Stack: A grammar for reading potential, broken into four observable layers: cognitive flexibility, learning agility, exploratory curiosity, and the capacity to unlearn. The four layers are not ranked; their relevance varies by zone. It produces no score — it makes legible the dynamic by which an employee learns and reconfigures.
See chapter 6
Psychological safety: The organizational condition in which an employee can take initiative, admit an error, or ask a question without fear of sanction. A concept formulated by Amy Edmondson at Harvard in 1999. Required in Laboratory zones, where experimentation is central. Dynamic reading systems have to be built so they do not destroy it.
See chapter 8
Radar: The technical device for dynamic reading of learning trajectories. It draws on multiple signals — mobility, learning usage, project history, documentation produced, cross-validations — to surface profiles that classic systems miss. The radar is not a judge: it is a partial signal, whose legitimacy depends on its safeguards.
See chapter 10
Reverse People Review: A partial inversion of the classic talent-review logic. The starting question is no longer only "who is ready to move up?", but also "who shortens other people's adoption time?", "who is central to spreading knowledge?", and "who looks average in the old system but turns out to be decisive in the new one?". It runs alongside the classic people review, not in place of it.
See chapter 8
Safeguard: The role a veteran expert can take in a hybrid team deployed on the Front Line. They bring their technical depth not as the project's main carrier, but as the last line of verification on the decisions the reconfiguration profiles produce. The role is what preserves the value of expertise.
See chapter 8
SBO: Skills-Based Organization. An organizational architecture that reads the organization through the stock of skills it holds. It has structured HR practice for roughly thirty years and remains relevant in Fortress zones. It becomes insufficient — without becoming obsolete — wherever reconfiguration speed creates more value than stable depth.
See chapter 2
Stabilization plateaus: Periods of consolidation during which management bans any change to work procedures, so that teams can absorb recent gains. A concept tied to allostatic load. A protective device against the glorification of technological chaos and the pile-up of adoption emergencies with no prior digestion.
See chapter 15
System of Record: The deepest layer of the HRIS, custodian of the contractual data: employment contracts, payroll, classification grids, seniority. In the book's model this layer is untouchable. It is the sovereign domain of traditional HR, and it must not be affected by the dynamic reading systems.
See chapter 8
Systematic human oversight: The second pillar of the protection doctrine. No consequential HR decision is made on the basis of an algorithmic reading without documented human validation. The machine proposes; the human disposes and signs. This pillar puts into practice the EU AI Act requirements (articles 6 and 9) on high-risk HR systems.
See chapter 13
Time-to-Skill: A measure of the speed at which an employee or a team takes ownership of a new skill scope. It is read through several observables: time to acceptable autonomy, level of support needed, transferability, sustained contribution, and the acceleration effect on others. It distinguishes individual Time-to-Skill from collective Time-to-Skill. Not a score — a reading of trajectory.
See chapter 7
Unforeseen-events leader: A role in a hybrid team. A profile able to arbitrate, in real time, the tensions between Fortress experts and fast learners; to prioritize projects; to hold against contradictory pressures; and to produce decisions that hold up before the sponsor. Identified through track records on projects under strain, and through cross-validation by peers.
See chapter 14

A

B

C

D

E

F

H

K

L

N

O

P

R

S

T

U