The hidden cost of inefficiency in project teams — and where AI fits in
- Athena P3M
- Feb 19
- 4 min read
Every project team wants to deliver on time, on budget and to scope. But the data tells a different story.
Recent research paints a stark picture of how project teams actually spend their time. According to the Wellingtone State of Project Management report, 50% of project workers waste at least one full day every month manually sifting through data and compiling reports. That's not delivery — it's admin. Multiply that across a programme with dozens of people and the lost productive hours are staggering.
It gets worse. 80% of organisations report spending at least half their time on rework — revisiting and correcting work that should have been right first time. In construction specifically, McKinsey has estimated that global inefficiencies cost the sector $1.6 trillion annually, with budget overruns ranging from 20% to 45%. And in the UK, 51% of construction professionals see budgets increase by 11–20% above the original estimate during delivery — driven largely by scope creep, change orders and design errors.
Where does the time actually go?
The patterns are familiar to anyone who's worked on a complex programme:
Manual reporting and data collation. Teams pulling numbers from disconnected spreadsheets, reformatting slides, chasing updates. 42% of project managers still spend a day or more each month manually compiling reports.
Rework from poor communication. 70% of professionals report that communication challenges within their organisation have led to wasted time. On a multi-stakeholder programme in defence or infrastructure, that's not just frustrating — it's expensive and potentially dangerous.
Inconsistent controls and governance. Only 48% of organisations consistently baseline their project schedules, and just 64% always or mostly engage in risk management. Without these fundamentals, teams are flying blind — and correcting course later always costs more.
Lack of real-time visibility. Half of all organisations still don't have access to real-time KPIs. Leaders are making decisions on data that's already out of date.
What does this mean for the future?
Project failure rates have actually risen — from 12% in 2025 to 13% in 2026 — despite growing investment in tools and methodologies. That tells us something important: buying software alone doesn't fix the problem. The issue is how teams work, how information flows and how decisions get made.
In critical environments like defence, energy, nuclear and infrastructure, these inefficiencies carry outsized risk. A delayed report isn't just an inconvenience — it can mean a missed risk, a late intervention, or a safety issue that should have been caught weeks earlier.
Where AI genuinely helps
This is where AI is starting to make a tangible difference — not by replacing project professionals, but by removing the friction that stops them doing their real job.
The Association for Project Management's latest survey of 500 UK business leaders found that 44% report AI is already having a positive impact on their bottom line, and 43% see a direct boost to productivity. LSE research goes further: employees using AI are saving an average of 7.5 hours per week — the equivalent of a full working day — worth around £14,000 per person per year in productivity gains.
In project teams specifically, AI is proving most useful in three areas:
Automated reporting and data synthesis — turning hours of manual collation into minutes, freeing teams to focus on analysis and decisions rather than formatting.
Predictive risk identification — surfacing schedule-risk patterns and cost trends before they show up in monthly reports. Early research suggests AI-enabled tools can reduce project failure rates by up to 35% and timeline deviations by 22%.
Scenario modelling — testing "what-if" options in minutes rather than weeks, so leaders can make better-informed trade-offs on live programmes.
But there's an important caveat. A recent study from Microsoft Research and Carnegie Mellon found that as employees grow more confident in AI's ability to execute tasks, their application of critical thinking diminishes. AI is a powerful lever — but only when paired with experienced professionals who know when to trust the output and when to challenge it.
A view from the ground
Tyler Skerton, Director, Athena P3M:
"The biggest inefficiency I see on complex programmes isn't a lack of data — it's too much data and not enough insight. Teams spend their week assembling information instead of acting on it. AI doesn't replace the judgement you need in a nuclear-regulated or safety-critical environment, but it does clear the path so experienced people can focus on the decisions that actually matter. The organisations getting this right are the ones treating AI as a controls tool, not a novelty."
Matt Giddings, Associate, Athena P3M:
"On the programmes I've worked across in defence and infrastructure, the amount of time lost to manual reporting and chasing updates is enormous. When you're managing hundreds of stakeholders across multiple organisations, even small inefficiencies compound fast. What excites me about AI in this space is the potential to give project managers their time back — not to do less, but to spend that time on risk, quality and the conversations that keep programmes on track."
The bottom line
The construction and infrastructure sectors are at a crossroads. Investment is growing, programmes are getting more complex, and the margin for error is shrinking. Yet project teams are still losing days every month to manual processes, rework and disconnected systems.
AI won't fix a weak delivery culture. But in the hands of capable project professionals, it can remove the friction that holds good teams back — and that's where the real efficiency gains lie.
We'll be exploring this theme further in our next Bridging the G_P markets report.
Comments