Resilience and Antifragility in the Management
of Critical Projects
FEATURED PAPER
By Luca Paolo Giuseppe Prinzio
Turin, Italy
Abstract
In critical systems built on cloud infrastructure and artificial intelligence, the unforeseen is not an exception to be eliminated but a structural condition. Technology evolves faster than the culture, governance and operational capacity that ought to govern it: between what a project plans and what the system actually does there remains an irreducible margin, and it is within that margin that incidents emerge. This article argues that, in such contexts, the objective of project management can no longer be limited to avoiding failure: it must also be to design the project-system so that it withstands and recovers when an unforeseen failure occurs, and so that the organisation learns from it.
The article distinguishes between robustness, resilience and antifragility, and applies the distinction to project practice while explicitly recognising its limits: in critical services and in Public Administration the realistic objective is resilience, whereas antifragility properly concerns processes, teams and learning, not the delivered service. It identifies the dimensions that determine whether a project withstands the shock — team cohesion, human sensors, the map of breaking points, decision-making under stress — and proposes operational practices drawn from reliability engineering: chaos engineering applied to management, cognitive redundancy, the error budget, the blameless postmortem. It closes by discussing the ambivalent role of artificial intelligence, which widens the gap between technology and culture and is itself a source of the unforeseen, but which can also strengthen the early detection of anomalies.
Keywords: project resilience, antifragility, chaos engineering, cognitive redundancy, error budget, Safety-I and Safety-II, high reliability organisations, requisite variety, critical systems, cloud, artificial intelligence, Public Administration.
- Introduction: The Unforeseen as a Structural Condition
A critical digital service, hosted on cloud infrastructure, had been carefully designed. There was a project plan; the tests had been passed; runbooks, a governance model and continuity procedures were all in place. Each component, taken individually, had been verified. Then, one night, the service degraded in a way no one had foreseen: an automatic update to a library, combined with an anomalous traffic peak and the silent degradation of an external dependency, produced a behaviour that no document contemplated. None of those three events, in isolation, would have been a problem. Their combination was. The recovery, in the end, did not depend on the quality of the plan — the plan did not foresee that scenario — but on the ability of a few people to improvise well, under pressure, reconstructing on the fly what was happening.
This scene is not an exceptional incident. It is the ordinary way in which complex systems fail. Many traditional plan-and-control interpretations of project management implicitly assume that, with sufficient planning, risk analysis and control, failure can be prevented. Contemporary standards are considerably more nuanced — the seventh edition of the PMBOK® Guide devotes a performance domain to uncertainty, and agile approaches are built around adaptation — but the assumption survives in organisational practice, where governance is still frequently designed as though the future were enumerable. In critical systems built on cloud infrastructure and artificial intelligence, that assumption does not hold. Not because planning is useless, but because there is a class of events that no plan can anticipate: unforeseen interactions between components, under conditions that no one has specified, at a speed that no control can match…
More…
To read entire paper, click here
How to cite this article: Prinzio, L. P. G. (2026). Designing for Failure: Resilience and Antifragility in the Management of Critical Projects, featured paper, PM World Journal, Vol. XV, Issue X, October 2026. Available online at https://pmworldjournal.com/wp-content/uploads/2026/10/pmwj169-Oct2026-Prinzio-Designing-for-Failure.pdf
About the Author

Luca Paolo Giuseppe Prinzio
Turin, Italy
![]()
Luca Paolo Giuseppe Prinzio is a certified Project Manager and Database Administrator at CSI Piemonte in Turin, Italy, where he participates in complex projects on cloud and security. For over twenty years he has worked in the ICT world and carries out teaching and consulting activities in the field of Project Management. He can be contacted at lprinzio@gmail.com and linkedin.com/in/lprinzio








