The construction industry has long been associated with delays, cost overruns and fragmented ways of working. Despite managing trillions of dollars in global capital expenditure, it remains one of the least digitised sectors.
According to Klemensas Mecejus, executive director and head of artificial intelligence (AI) advisory at ai71, AI is now at a point where it can fundamentally change that reality, not through isolated pilots, but through enterprise-scale transformation.
Mecejus’ journey into AI began not in research labs, but in large-scale analytics and data transformation programmes. Early on, he saw the potential of machine learning to improve real operational decision-making inside complex organisations. “The real challenge wasn’t building impressive models,” he says, “it was making AI work across an entire enterprise, embedding it into how decisions are made.”
That focus ultimately led him to ai71 and its construction-focused platform SuperHive, which aims to turn highly fragmented project data into actionable intelligence across the full lifecycle of design, delivery and operations.
From reactive to predictive construction
According to Mecejus, construction’s problems are well understood but rarely quantified so starkly. Manual, low-value activities consume up to 40% of project time. Fragmented systems between developers, contractors and regulators can result in 20% rework. Unstructured data drives a 25% productivity loss, while limited real-time visibility and lack of predictive insight contribute to persistent cost and budget overruns.
“AI is shifting construction from reactive management to predictive and proactive decision-making,” he says. “That change is far more significant than most people realise.”
In planning and design, AI-driven optimisation and simulation can identify risks far earlier than traditional approaches. Mecejus points to solutions that have reduced pre-concept master plan assessments from months to minutes, as well as automated CAD-to-BIM conversion tools that eliminate weeks of manual work per project.
During delivery, AI-enabled monitoring is already transforming oversight and compliance. SuperHive’s Analyze Studio automates regulatory checks that previously took weeks, reducing permit approval times from one month to around one minute while improving accuracy from 42% to 99%.
Elsewhere, ai71 has deployed satellite imagery, drones and IoT-based monitoring across more than 100,000 residential units and 150 towers, cutting idle construction time by 20% and significantly reducing safety violations.
In operations, Mecejus argues that AI finally allows project data to become a long-term strategic asset rather than a static archive. Predictive maintenance, digital twins and continuous asset performance monitoring are replacing periodic, manual inspections with real-time intelligence.
Why the Middle East can move faster
Mecejus believes the Middle East is uniquely positioned to leapfrog other regions in AI-driven infrastructure development. Strong government backing for digital transformation, access to capital and a willingness to deploy technology at scale all play a role. Crucially, many organisations are not constrained by decades of legacy systems.
“We work with real-estate players managing US$20bn-50bn in assets,” he says. “They can adopt modern AI platforms far more quickly because they don’t carry the same technical debt you see elsewhere.”
“AI is shifting construction from reactive management to predictive and proactive decision-making”
Klemensas Mecejus, ai71
Large, centrally managed infrastructure programmes also make it easier to standardise data and scale AI across portfolios. The company has already developed AI-powered digital twins for entire cities in the region, supporting large-scale planning by simulating the impact of development on neighbourhoods and infrastructure.
However, Mecejus cautions that speed must be balanced with governance. Without clear operating models, talent development and accountability, rapid AI adoption can stall. “The organisations seeing the biggest returns are those that treat AI as core infrastructure, not experimentation,” he adds.
In his advisory work, Mecejus says successful AI programmes always begin with clearly defined business outcomes. Too many initiatives fail because they start with technology rather than value creation. Data readiness is particularly critical in construction, where up to 80% of information is unstructured and scattered across disconnected systems.
“AI can’t fix broken workflows,” he notes. “If tender documents, regulatory submissions and project data are all fragmented, no algorithm will save you.”
Equally underestimated are organisational hurdles such as unclear ownership of AI outcomes and poor change management. Mecejus highlights a B2C residential sales platform built by his teams that reduced the buying process from weeks to minutes while cutting sales staffing by 80%. The key to adoption, he says, was prioritising user experience rather than the AI itself.
The next frontier: agent-based AI
Looking ahead, Mecejus sees the biggest untapped opportunity in turning project-level insights into reusable, enterprise-wide intelligence. AI-driven compliance and quality assurance, particularly for large public infrastructure programmes, could save the industry billions by replacing inconsistent, manual reviews with standardised, high-accuracy automation.
Over the next two to three years, he expects AI to be embedded across the entire construction lifecycle, supported by agent-based systems that actively assist engineers, project managers and regulators.
“Imagine asking your construction platform which projects are most at risk next quarter and getting an immediate answer with evidence and actions already underway,” he says. “That’s where the industry is heading.”