Singapore's construction industry is gearing up for its strongest year since before the pandemic, with demand forecast to reach between $47 billion and $53 billion in 2026.

Yet speakers at the BuildSG Lead Summit noted that the built environment sector faces mounting headwinds, and long-term competitiveness will depend on whether firms can move from AI experimentation to scaled, enterprise-wide deployment.

Organised by the Building and Construction Authority (BCA), this year’s event was themed "Driving value: Leadership in the age of intelligent built environment".

Read also: 'Errant' housing developers will be disqualified from government land sales, barred from selling projects for up to five years

It covered topics ranging from AI governance and data readiness to talent development and collaborative contracting, as the industry pushes to embed technology as a strategic capability.

Held on April 30, the summit opened with an address by Minister for National Development, Chee Hong Tat, who announced three pro-enterprise measures.

These include a streamlined approval process for developers and project teams adopting the Kit-of-Parts approach; the progressive removal of the overseas testing requirement for new construction workers; and a new annual rating survey for consultants to provide feedback on government agencies.

Lean construction, collaboration, robotics

Following the announcements, Derek Tan, BCA CEO (designate), laid out the agency's industry workplan for the year, amid a buoyant demand outlook for the construction sector alongside rising uncertainties.

Pointing to the industry's "very strong" rebound after the pandemic, Tan said that total construction demand in Singapore is expected to reach up to $53 billion this year, nearly $17 billion above pre-Covid levels.

At the same time, firms will need to navigate challenges such as geopolitical tensions, manpower constraints and rising material costs.

Read also: The AI shift and the Island: How AI-driven uncertainty elsewhere could fuel residential demand in Singapore

To build resilience and sustain growth, firms should focus their efforts in three areas — lean construction, collaborative contracting, as well as robotics, automation and AI, Tan said.

Before outlining the road ahead, he noted progress made in 2025, including the significant take-up of Design for Manufacturing and Assembly (DfMA) and the Integrated Digital Delivery (IDD) framework.

DfMA involves prefabricating building components off-site in a controlled environment, before assembling them on-site, while IDD uses digital technologies to connect all stakeholders across a building’s lifecycle. Adoption rates for both have increased to 76%, surpassing the 70% target for 2025.

On lean construction, Tan said this approach will enable the next level of productivity savings, building on the foundation laid by DfMA and IDD. Studies show that 57% of productive time in construction is wasted, with downtime, rework and poor coordination among the common culprits of inefficiency.

The lean construction model aims to minimise waste, maximise value and emphasise collaboration between teams.

Early results from firms already on the lean journey have been encouraging. In the design and construction of Ang Mo Kio MRT Station and tunnels under Cross Island Line Phase 1, a critical traffic diversion milestone was completed two weeks ahead of schedule. Site bore piling activity was also sped up by 12 days, and project profitability was improved.

Read also: World's biggest construction startup competition to debut Asia Pacific pitch day in Singapore

The second focus area under the BCA-industry workplan is collaborative contracting, which Tan described as "a better way to share risk and reward" as compared to conventional lump-sum contracts — especially for large, complex projects.

Collaborative contracting allows for proactive project management, risk sharing and joint problem-solving. This can translate to better coordination, more certainty on timeline and cost, and fewer disputes, he added.

For example, the LyndenWoods condo project along Science Park Drive was the first private-sector project using the NEC4 collaborative contract in Singapore, and demonstrated equitable cost-sharing of risks.

As the NEC4 is a standard contract form used internationally, BCA has developed additional clauses for use in SIngapore across engineering, construction, and facilities management projects.

On robotics, automation and AI, Tan urged firms to start building capabilities now to seize future opportunities.

At Parc Meadow @ Tengah, a painting subcontractor deployed a fleet of painting robots for the Build-To-Order project, which halved the manpower needed. Instead of the usual two-men painting team, only one worker was required to manage two robots, and two robots also could paint up to eight five-room HDB flats per day.

These priorities — spanning productivity, collaboration and technology adoption — were echoed by industry leaders in the presentations that followed.

Why construction AI fails and what to fix first

Speakers shone the spotlight on how AI and digital transformation are being applied in the built environment.

Boustead Singapore’s senior vice-president, group technology office, Muhammad Khalil Shaiful Bahari, shared how the engineering and technology group is executing its AI roadmap in practice, and the lessons learnt from failed pilots.

“We tried seven AI tools across our projects in the last two years. Three are still running. The other four taught us more about enterprise AI than any conference I’ve attended,” he said.

When construction AI tools fail and join the “proof-of-concept graveyard”, it is more often because the organisation was not ready and had been skipping the foundation work — and less because the technology was immature, Khalil pointed out.

Data chaos is one way this lack of organisational readiness shows up. AI, being a pattern-recognition engine, needs clean structured data, but the construction industry tends to feed it fragmented files across emails, WhatsApp and multiple folders per project.

Workflows and processes should also be fixed first before they can be augmented with AI. “The most dangerous AI deployment is one inserted into a broken process, as it automates the dysfunction and scales it,” Khalil said.

Another reason construction AI fails is what he called the “people gap”, where technology adoption is more a challenge of change management in an organisation.

In applying these learnings, Boustead has put in place a decision framework that prioritises data, processes and governance. For instance, it maps every workflow that AI is expected to touch and eliminates bottlenecks before automating. Each initiative is also given a readiness score based on its impact, feasibility and data.

Having a human in the loop is a non-negotiable for Boustead. Every AI-assisted output carries a clear record of review by a person, and no output goes to a client without a named engineer’s sign-off. Given the risk profile of construction, human review is also mandatory for any automation of safety-critical structural or fire compliance decisions.

Data first, and bringing people along

Matt Warren from real estate developer and construction business Lendlease spoke about bridging the gap between design intent and built outcomes, and how design intelligence combined with AI-enabled data analysis can drive better decision-making.

Likewise, he stressed that data should come first, and AI second. “Achieve these three things for your data: simple, standardised and siloed,” said Warren, who is head of digital at Lendlease, Europe.

People are another key piece in the puzzle, with leadership and organisational alignment crucial in enabling transformation. To close the knowledge and experience gap in AI, he suggested organisations find four key traits in their people: empathy in leaders, collaboration, diversity of thought, and a focus on results.

Empathetic leaders are essential, as they will need to understand that fear is often the single biggest blocker to adoption. Diversity of thought and innovation are also key, as “your domain-specific language model will learn from you and the questions your team asks”, Warren said.

He urged leaders to prepare and empower their employees. “The greatest risk to our industry is not AI; it is leaving our people behind,” Warren added.

Making AI work across the value chain

Google Deepmind’s director of AI impact, Ramine Tinati, shared the latest AI advancements and pointed out that AI can help reimagine the built-environment value chain — such as by forecasting fluctuations in material prices, predicting the weather’s impact on thermal performance, providing robotic support for construction site tasks, and using sensor data to optimise space utilisation in properties.

He recommended a four-step approach for firms to build AI responsibly, beginning with a data audit to assess data hygiene before any AI investment, followed by a "high-value" pilot to prove the return on investment quickly and then scaling up and outwards.

Next, the company should shift its focus to the talent strategy, moving from execution to “AI orchestration” to empower engineers and architects to act as “AI co-pilots”, Tinati said. Finally, the team will need to break silos by partnering with others in the ecosystem and value chain.

In another presentation, Arup’s David Moran talked about moving from project-level innovation to scalable digital delivery at the enterprise level.

He also highlighted the built-environment consultancy’s “total design” approach, which integrates AI, parametric design and automated workflows.

This approach shows how earlier alignment across disciplines can accelerate design timelines, reduce embodied carbon and minimise downstream rework, according to Moran, who leads the digital business for Arup’s Asia Pacific (Apac) region.

Turning AI ambition and hype into results

The summit’s panel discussion distilled many of these themes and focused on how firms can move beyond AI experimentation and pilots to apply the technologies in ways that deliver real value.

A key message from the session — moderated by Kok Su-Ming, managing director of BCA Academy — was the need for a clear vision and purpose, strong leadership and the right organisational culture.

"Using AI for the sake of using AI can be very, very dangerous," cautioned Haresh Khoobchandani, vice-president for Apac and Japan at Autodesk, a design, engineering and construction software firm.

Stressing that technology is an enabler but should never be the starting point, he said firms must first define the problems they are trying to solve, and understand that data is a business priority.

Echoing this, Moran from Arup brought up the importance of practical application of AI. Firms should assess whether the outputs are actually useful in real project contexts, while continuing to explore and experiment with the technology.

"I’m trying to ban PowerPoint slides about how powerful AI is going to be in the future," he said half-jokingly, drawing laughter from the audience. "I’d rather see it in action and then have a reality check … Is it actually any good? Is what you’re delivering going to [help with] your Friday deadline or is it just some AI magic?"

To ensure AI delivers real-world outcomes, the panellists pointed to data, governance and broader adoption.

They called for better visibility and flow of data across the value chain, as well as clear accountability frameworks as AI becomes more embedded in workflows.

Summing up the session, Kok said that applying AI is "not just about doing things better", but about deciding "what is worth doing differently".

This shift will be critical as the built environment sector moves beyond pilots and isolated use cases to wider, real-world adoption at scale.

For more news and analysis, read our weekly e-paper. Prefer a print copy? Get it delivered to your home every Monday.