Every knowledge worker recognises the phenomenon. A project sits frozen on a digital task board, its status unchanged for weeks. The internal discussion channel has gone quiet. Nobody genuinely believes it will ever be completed, yet it remains officially “active”. Welcome to the world of zombie projects — initiatives that are neither alive enough to progress nor decisively shut down.
The scale of the problem is striking. Research by Atlassian surveying thousands of office workers globally suggests that nearly half began 2026 weighed down by such undead initiatives. More troubling, over 90 per cent report tangible consequences: stress from cluttered workloads, declining productivity, and growing concerns about burnout.
These are not minor irritations. At a time when executives insist their organisations must move faster simply to remain competitive, zombie projects quietly consume time, attention and cognitive energy that should be directed towards innovation, execution and AI adoption. Yet despite widespread recognition of their cost, such projects persist — suspended in a peculiar limbo between abandonment and revival.
Consider a typical case: a customer experience redesign initiative launched with fanfare in March 2024. The project consumed three months of intensive work before stalling when priorities shifted. Eighteen months later, it remains officially “in progress” on project dashboards, though the original sponsor has moved to another company, the budget has been quietly reallocated, and market conditions have evolved substantially. Everyone knows it will never be completed, yet no one has formally closed it. The project exists in a state of suspended animation — neither alive nor officially dead.
Zombie projects don’t fail loudly — they rot quietly, draining time, morale and attention
Why the undead persist
If zombie projects drain morale and productivity, why are they so rarely killed off? The answer lies less in individual behaviour and more in structural weaknesses common to modern organisations.
A significant proportion of employees report reluctance to abandon projects for fear of how it will be perceived. Ending an initiative is often interpreted as failure or insufficient effort, even when underlying conditions have changed materially. The safer option, psychologically, is to leave projects dormant rather than confront the optics of closure.
Compounding this is what researchers describe as a “decision gap”. In many organisations, no one has explicit authority to declare a project finished. Ownership of initiation is clear; ownership of termination is not. In the absence of defined decision rights, projects continue by default.
The sunk cost fallacy reinforces this inertia. Teams grow attached to initiatives that have absorbed months of effort, resources and emotional investment. Acknowledging that further work will not justify past expenditure is uncomfortable. Leaving projects technically alive preserves the illusion that the effort may yet be redeemed.
Zombie projects also reflect weak strategic focus. Initiatives that lack a clear connection to organisational priorities gradually drift down the agenda. They are never resourced sufficiently to progress, yet never challenged strongly enough to be discontinued.
Timing plays its own role. Research shows that projects most commonly stall in mid-December, when teams defer decisions to the new year. January then arrives already burdened with commitments few genuinely intend to complete.
When no one owns the decision to stop, unfinished work becomes organisational clutter.
Enter artificial intelligence
Could artificial intelligence help resolve a problem that is fundamentally human? Survey data suggests cautious optimism, though with notable variation by geography. Globally, a majority of workers believe AI tools could help guide decisions on whether to revive or retire stalled initiatives — a sentiment particularly strong in India.
The appeal of AI in this context lies in its emotional detachment. Unlike human managers, AI systems are not invested in past decisions or reputational outcomes. They can assess projects based on current strategic relevance, resource availability and feasibility, rather than historical effort.
Workers identify several practical applications. Many want AI to act as a comprehensive summariser, surfacing full project context so teams can quickly understand what exists and what does not. Others value realistic time and effort estimates grounded in actual workloads, rather than optimistic assumptions. Still others see value in AI extracting insights, decisions and action items from archived emails, documents and internal communications.
These capabilities address a persistent inefficiency. Large organisations lose billions of hours annually to information hunting and context reconstruction. AI that can surface relevant material quickly may recover significant amounts of productive time — and reduce the friction that keeps zombie projects alive by default.
AI won’t kill zombie projects — but it may finally force leaders to look them in the eye.
Collaboration, not execution
Positioning AI as an executioner, however, risks overstating both its capabilities and the nature of the problem. The core issues that sustain zombie projects — unclear authority, misaligned incentives and weak strategic clarity — are organisational, not technical.
AI can surface information, identify patterns and flag inconsistencies. It cannot decide what matters. Nor can it resolve political sensitivities, personal attachments or risk aversion that often keep unviable initiatives alive.
The more realistic promise lies in human–AI collaboration. AI can synthesise project histories, highlight dependencies, estimate resource requirements and identify strategic drift. Humans must still make the judgment calls about continuation, reprioritisation or closure.
Consider how this might work in practice with that stalled customer experience redesign. An AI system could instantly surface the original business case, document how market conditions have shifted since launch, note that the project champion left the company eight months ago, identify that three of the five core team members have moved to different roles, and estimate that revival would require 400 hours of specialist work. It might also highlight that customer priorities identified in focus groups eighteen months ago differ markedly from current feedback. But a human must still decide whether those 400 hours are better spent elsewhere, whether the original problem remains relevant, and how to communicate the closure decision to stakeholders.
Used well, this division of labour plays to respective strengths. Machines process information at scale. Humans navigate strategy, trade-offs and organisational dynamics.
One practical application is automated detection. AI systems could flag projects displaying classic zombie signals — prolonged inactivity, repeated deadline extensions, dormant communication channels or minimal resource usage — and surface them for human review. Another lies in reframing. Rather than presenting decisions as project “termination”, AI could frame them as strategic reallocation, reducing emotional resistance whilst leaving the decision itself unchanged.
The real cost of zombie projects isn’t wasted effort, it’s delayed focus on what actually matters.
Limits worth acknowledging
Several caveats remain. AI’s usefulness depends heavily on data quality. Projects documented inconsistently or conducted largely through informal conversations may resist algorithmic analysis.
Moreover, better information does not automatically resolve organisational dysfunction. Even when everyone recognises that a project should end, political considerations and risk avoidance can override logic. AI does not neutralise power dynamics.
There is also a longer-term risk. Excessive reliance on algorithmic guidance may erode human judgment and accountability, creating new dependencies whilst solving old inefficiencies.
If a project survives without purpose, the problem isn’t execution — it’s decision-making.
Tools enable; culture decides
Zombie projects ultimately reveal more about organisational culture than technological capability. They thrive where accountability is diffuse, strategic priorities are unclear and decisive action carries perceived risk.
AI can help surface information, reduce administrative drag and make trade-offs more visible. But eliminating zombie projects requires something less technical and more difficult: clarity about who decides, what matters and when something is allowed to end.
Organisations that succeed will not be those with the most advanced AI tools, but those that use technology to support clearer thinking — whilst addressing the cultural conditions that allowed undead initiatives to persist in the first place.
Zombie projects, then, are a useful diagnostic. Where they proliferate, deeper issues of decision-making and accountability are rarely far behind. The question is not whether AI can kill them, but whether leaders are willing to use this moment to confront what kept them alive.