Mid-size construction firms in Ireland and the UK are pulling ahead by adopting AI tools early, boosting productivity by up to 20%, according to a new report from the AI Institute. In contrast, larger and smaller firms are falling behind, with fewer than a third offering AI literacy training and only one in five having formal AI policies.

The survey, covering 29 senior representatives across construction, architecture, engineering, and quantity surveying, found widespread ‘shadow use’ of AI, exposing firms to cybersecurity risks and inefficiencies. Professionals reported fragmented and siloed data, minimal automation, and limited AI knowledge, with 46% citing lack of training as a barrier to harnessing AI’s potential.

Mid-size companies are leveraging their agility to implement AI strategies from the ground up, establishing policies, training staff, and identifying automatable processes. Larger firms, despite investing heavily, are hampered by insufficient policy frameworks, poor change management, and inadequate staff training. Cultural resistance to technology, particularly among SMEs, remains a major obstacle, even as these firms are often best positioned to benefit.

The AI Institute emphasises that effective AI adoption starts with mandatory staff training, formalised AI policies, and standardised data practices. Establishing homogenous data protocols across projects will allow AI tools to operate reliably, minimise compliance risks, and deliver measurable business impact.

The report notes that commercial large language models (LLMs) are already widely used in the sector, often without compliance safeguards. Tailored AI solutions built around a firm’s specific needs, combined with strong foundational policies, are crucial for translating AI investments into sustained productivity gains.

Mid-size firms are demonstrating that early, structured AI adoption creates a clear competitive edge, offering lessons for any business navigating the integration of AI into traditional industries.

Discover how your firm can translate AI potential into real-world results by exploring the full report.