AI Insights: Omnix International’s Walid Gomaa 

The engineering and construction industries have built their reputation on precision rather than experimentation. Decisions are grounded in proven methods, timelines are unforgiving, and the standards for delivery leave little room for uncertainty. Which is exactly what makes the pace and depth of AI adoption across the built environment sector so telling, and the nature of that adoption worth examining carefully. 

Across asset planning, intelligent design within BIM platforms, real-time site monitoring, predictive maintenance, and autonomous building operations, the conversation has moved on from feasibility. The organisations shaping this moment are not doing so by running smarter pilots. They are doing so by redesigning how they operate. 

The question that defined the previous cycle, whether AI could deliver real value in complex, asset-heavy environments, has been answered. What has replaced it is harder and more consequential: whether organisations have the structural architecture to sustain AI beyond the pilot stage and run it as a continuous operational function.  

“AI has moved from isolated pilots to becoming a structural pillar of the engineering and built environment sectors,” says Walid Gomaa, CEO, Omnix International. “Real AI at scale means intelligence is no longer experimental, it is embedded into live workflows, connected to operational data, and governed for consistent, repeatable execution. The real differentiator is depth. AI is now embedded within Common Data Environments, digital twins, and cost-control systems.” 

The operationalisation gap 

The gap between a successful pilot and a production-grade AI capability is where enormous amounts of value currently disappear, and the reasons are rarely technical.  

The organisations making genuine progress share structural characteristics that have less to do with which models they are running and more to do with how their data, talent, and operations are organised. They prioritise unified, well-governed data architectures, because fragmented or siloed data is not a foundation on which AI can scale. They invest seriously in MLOps capabilities, the infrastructure, processes, and talent required to deploy, monitor, and continuously refine models once they leave controlled environments.  

“Without this, models that perform well in pilots degrade in production, undermining trust and value,” explains Gomaa. “Programme managers, engineers, and asset operators who understand AI’s capabilities are critical to translating insights into decisions. These organisations treat AI as an operational capability, not a standalone initiative. They embed AI into operations rather than isolating it within innovation teams.” 

The convergence of AI with IoT is where this operational logic becomes most visible in the physical world. Sensor networks, continuous data streams, and real-time processing are transforming infrastructure from passive assets into systems that actively participate in their own management. During project delivery, IoT enables real-time tracking of safety, materials, and progress. In operations, those same networks support continuous optimisation of building systems, from HVAC performance to occupancy management.  

“The most tangible impact is predictive maintenance, reducing costs by 20 to 40 percent, extending asset life, and minimising downtime. This integration ensures infrastructure is not just built, but continuously optimised, actively participating in its own performance and lifecycle management,” says Gomaa. 

Underpinning all of this is an infrastructure question that many organisations are only now confronting at the scale AI demands. Traditional IT environments were not designed for the compute density, memory bandwidth, and storage performance that serious AI workloads require, and the gap is becoming increasingly difficult to manage around.  

The response across the sector is not a wholesale migration to cloud, but a deliberate hybrid strategy calibrated to workload requirements. “Cloud platforms remain critical for training models and experimentation, offering elastic scale and access to advanced accelerators,” he says. 

“However, latency-sensitive and mission-critical applications increasingly rely on edge and on-premises infrastructure for real-time processing and resilience. Rather than treating compute, cloud, and infrastructure as separate layers, organisations are integrating them into a unified AI foundation. This approach balances performance, scalability, cost, and security, ensuring the infrastructure aligns with operational demands.” 

The trust architecture 

As AI becomes embedded in systems that interact directly with physical environments, the consequences of a governance failure extend well beyond data integrity, and this is where the sector’s thinking is evolving most rapidly. “AI now interacts with physical systems, meaning vulnerabilities can impact safety and critical infrastructure, not just data,” says Gomaa. “Organisations are adopting integrated IT/OT security frameworks based on zero-trust principles, ensuring continuous verification across users, devices, and systems. At the same time, AI governance is becoming more structured, with clear accountability, oversight, and escalation mechanisms. Leading organisations treat governance not as a constraint, but as an enabler of safe, scalable innovation, ensuring operational integrity while advancing AI adoption.” 

The GCC occupies a distinctive position in all of this, and not simply because of the scale of investment flowing through the region’s infrastructure pipeline. The structural advantage is the opportunity to design projects with AI embedded from inception, rather than retrofitting intelligence into legacy systems built around entirely different assumptions. “Large-scale projects are being designed with AI embedded from inception, redefining the baseline for execution,” Gomaa says.  

He adds: “Today, a credible AI strategy is no longer optional, it is fundamental to competing in a data-driven, autonomous delivery environment.”  

The leadership implications of this are ones Gomaa returns to with some consistency. The role of the engineering or IT leader is no longer defined by managing technology decisions, but by the ability to act on the intelligence those systems produce, under real operational conditions and at genuine scale. “The benchmark is no longer incremental improvement, but materially better outcomes achieved with fewer resources, lower risk, and faster responsiveness,” he says. “As operations become more data-driven, leaders must interpret and act on AI insights with confidence. Beyond technical expertise, success now requires strategic vision, organisational influence, and the judgment to balance human oversight with algorithmic recommendations.”