Annie Liang, co-founder of Billie Onsite, looks at why infrastructure projects are not kept on track.
In the United Kingdom, major infrastructure projects face persistent delays and cost overruns, with UK roads at 66% overruns and social infrastructure at 56%, surpassing European peers. Key cases illustrate: HS2's costs rose from £37.5 billion to £110 billion (193% overrun) with seven-plus years delayed; Hinkley Point C from £18 billion to £46 billion (155% overrun) and four years late; Elizabeth Line (Crossrail) from £14.8 billion to £18.9 billion (28% overrun) with 3.5 years postponed. These delays arise from prolonged coordination among stakeholders like governments, communities, regulators, and contractors in plan assessments and design changes. These pre-construction issues feed into onsite disarray, where changes cause rework, resource mismatches, supply disruptions, and productivity drops.
At the crux of the productivity drain there are two critical, self-inflicted dysfunctions embedded in the industry's culture: a costly cycle of rework driven by reactive compliance, design and quality controls, and a systemic failure of knowledge retention that forces every project to repeat the mistakes of the past.
The rework: Building It twice
One of the key drivers of construction's productivity drain is its deeply ingrained, reactive approach to compliance, design review and quality control. For too long, even within the integrated D&B (design and build) model where the contractor shoulders both design and construction responsibilities, accountability has been fragmented. Compliance, design and quality control is not treated as a continuous, internal responsibility for the project team but as a series of external pass/fail checkpoints. Oversight is effectively delegated to the design team, QA team, third-party inspectors and officials whose primary function is to find problems and defects after the work is done, institutionalising rework as a standard operating procedure.
This model carries an immense financial and operational price, directly contributing to stagnant productivity.
Financial drain
Rework accounts for 5-21% of a project's total value in UK construction, with infrastructure projects often at the higher end due to complexity.
Wasted labor
Up to 30% of all work on typical UK construction projects, including infrastructure, involves rework.
Schedule delays
Design-related rework contributes to average delays of 10-20% in major UK infrastructure, exacerbated by scope changes in rail projects.
Ironically, the root causes are largely preventable. Up to 80% of all rework can be traced directly back to design errors and changes - issues that a robust and continuous internal review process should identify long before a single worker is on site.
In rail projects, this manifests in errors from insufficient interdisciplinary checks or failure to incorporate site feedback, as seen in transport mega-projects where design weaknesses propagate into construction delays.
The industry could benefit from a fundamental shift from reactive assessments and inspections to proactive, internalised ownership. This means embedding disciplined systems aligned with UK standards, such as mandatory in-process verifications at key milestones (e.g., pre-installation checklists, verified by site superintendents) or real-time digital documentation via BIM tools for MEP installations before closure.
These interdisciplinary checks can be largely expedited by the use of AI. By empowering experts and project managers with the ability to quickly capture design and onsite issues by voice, images, and annotations on plans, and using AI to verify these issues against design documentation and building regulations, much can be communicated across various stakeholders at a much faster speed, and reports automated instantly.
The knowledge black hole: Repeating past mistakes
Compounding the rework drain is a parallel loss of institutional memory. The construction industry operates with a project-based mindset that treats each job as a disposable entity. When a project is completed, the teams disband, and the invaluable knowledge generated on-site - clever solutions, effective remediation strategies, hard-won insights - simply evaporates. This collective amnesia forces each new project team to reinvent the wheel, confronting the same predictable problems with no benefit from past experience.
This failure to learn from past lessons is one of the key reasons the industry has not seen the compounding productivity gains of manufacturing. Manufacturing's success is built on systematic process improvement - a Kaizen-like philosophy of continuous learning that is impossible without robust knowledge retention.
The data-rich, insight-poor industry
The industry is drowning in data but starved for wisdom.
There is a widespread recognition among construction firms that they have significant room to improve their data capabilities. On average, these firms actively analyse only one-third of the vast data they collect.
This inefficiency forces construction professionals to spend an average 11.5 hours per week - nearly a day and a half - simply searching for information, not analysing it. This waste is the direct result of an absence of knowledge management strategies, leaving critical information trapped in disconnected spreadsheets, email chains and inaccessible data silos.
An intelligent knowledge base
The antidote to institutional amnesia is the creation of a centralised, AI-powered knowledge management system as a core business function. This is not about creating another silo, it is about building a living, searchable library of corporate wisdom. Such an intelligent system would automatically capture, tag and categorise all critical project data, from inspection photos and rework logs to RFI resolutions and material substitution approvals.
The goal is to empower teams with on-demand intelligence. Imagine a project manager facing a challenging waterproofing detail. Instead of starting from scratch, they could instantly search the knowledge base to access the approved submittals, field-tested installation methods and quality inspection checklists from five previous projects that successfully solved the exact same issue. This transforms a history of isolated problems into a powerful, accessible, and actionable competitive advantage.
From lagging to leading
The story of the last half-century in construction has been one of stagnation, a narrative confirmed by data. This is an industry that too often builds things twice and learns things once, only to immediately forget them.
The path forward requires a radical cultural shift.
- The industry must move from a model of fragmented design review and inspection responsibility to one of proactive, internal ownership to finally build it right the first time.
- The industry must build intelligent knowledge management systems to stop solving the same problems over and over again.
The solutions to the construction productivity slump are unlikely to be found in blaming external forces. They are more likely to be found by looking inward, embracing radical internal accountability, and committing to digital and AI transformation.