Beyond the AI Hype: Reflections from the Bisnow AI & Technology Conference, New York
By Steve Cassells, Director, Neuron
Steve Cassells
Director
Neuron Director Steve Cassells recently spoke at the Bisnow AI & Technology Conference in New York, joining leaders from across real estate, construction, technology and academia to discuss the evolving role of artificial intelligence in the built environment.
The conference covered everything from AI adoption and implementation challenges through to competitive advantage, data strategy and emerging success stories.
What surprised Steve most was how much the discussion has shifted over the last 12 months. A year ago, many conversations focused on what AI might be capable of. Today, the focus is increasingly on implementation, governance, data quality and measurable business outcomes.
The industry appears to be moving beyond asking whether artificial intelligence is important and towards a far more practical question:
Where is AI genuinely creating value?
Data Quality Still Matters More Than AI
One of the strongest themes throughout both the panel discussions and the conversations that followed was that there is still no shortcut around data quality.
Despite the rapid advancement of AI capabilities, organisations continue to face a fundamental reality: if the underlying information is fragmented, inaccessible, poorly structured or simply doesn't exist, the outcomes will be limited regardless of how sophisticated the AI becomes.
The phrase "junk in, junk out" may feel dated, but it remains remarkably relevant.
Many organisations are eager to explore AI opportunities, yet the biggest challenge is often not choosing the right AI tool. It is establishing the information foundations necessary for those tools to operate effectively.
This is particularly true in property, construction and engineering, where information is created across multiple stakeholders, systems and project phases, and where context is often as important as the data itself.
The Future Is Not One Giant AI System
Another recurring discussion centred around what the future technology landscape is likely to look like.
There was broad agreement that the future is unlikely to consist of a single AI platform capable of managing every business process. Instead, a more probable outcome is an ecosystem of specialised tools, databases and intelligent workflows that exchange information between systems. In this environment, connected information becomes increasingly important.
The value is not simply in having access to AI. The value comes from creating environments where people, software and AI can work together effectively with greater context and understanding.
Businesses that invest in interoperability, connected data and open architectures today will likely be far better positioned to take advantage of future innovations.
AI Performs Best When It Has Strong Foundations
One observation that resonated strongly with us was the distinction between AI-enhanced products and AI-native products.
Across multiple industries, organisations are finding significant success applying AI to existing software platforms, workflows and knowledge bases. When AI is layered on top of established systems, it can dramatically improve productivity, automation, searchability and access to information. However, products that place AI at the centre of engineering or commercial decision-making processes face a significantly higher degree of scrutiny and risk.
Questions around accuracy, repeatability, transparency and accountability become increasingly important when users begin relying on outputs to make critical decisions. For many applications, being correct most of the time is not enough. This distinction is particularly important in engineering, construction and property development, where decisions often carry substantial financial, technical and safety implications.
Understanding the Sphere of Complexity
At Neuron, we often talk about a concept we call the Sphere of Complexity.
The outer layers of the sphere represent lower-complexity activities that are primarily information-processing tasks. These are the areas where AI is already delivering significant, measurable value. However, as AI moves deeper into the sphere towards the highly complex core, limitations begin to emerge. This is where fractures, errors and challenges become more apparent, as tasks increasingly require judgement, contextual understanding and the ability to navigate uncertainty.
Ultimately, the closer a task is to pure information processing, the more effectively AI performs today. The further a task moves towards judgement, trade-offs, uncertainty and accountability, the more essential human expertise becomes.
Solving Problems That Matter
Perhaps the most important takeaway from the conference was that technology alone does not create value. Regardless of how impressive an innovation may be, it ultimately needs to solve a problem that somebody is willing to pay to have solved.
The AI industry has experienced an extraordinary amount of attention, investment and marketing over the past several years. While this has accelerated innovation, it has also created unrealistic expectations in some areas.
Many of the biggest opportunities in construction over the next decade are likely to come from practical improvements rather than revolutionary breakthroughs.
Practical improvements such as:
Reducing costs through productivity improvements
Increasing the speed of project delivery
Improving coordination and communication
Making information easier to access
Reducing rework
Improving business operations
Supporting better decision-making
These outcomes may not generate headlines, but they have the potential to create enormous economic value across the industry.
Building for the Future
At Neuron, our long-term vision remains centred around creating cloud-based, API-driven feedback loops that enable better collaboration, faster decision-making and seamless information sharing across the built environment.
We believe the future will be increasingly connected.
Digital twins, interoperable software platforms, shared project intelligence and AI-enhanced workflows will continue to transform how projects are designed, delivered and operated.
Importantly, in the short to medium term, we do not see these technologies replacing the people responsible for designing and delivering buildings. Instead, they will augment human expertise, reduce friction, improve visibility and allow professionals to spend more time solving the complex problems where experience and judgement create the greatest value.
While much of the discussion around AI focuses on documentation automation, we believe one of the largest opportunities lies earlier in the project lifecycle.
By providing faster feedback on feasibility, cost, sustainability and design outcomes, technology has the potential to improve decision-making when it matters most.We refer to this as Feasibility-Led Design: using data and engineering insight to help teams make better decisions before problems become expensive to solve.
For an industry facing challenges around affordability, sustainability and productivity, this may ultimately prove to be one of the most valuable applications of technology.
Final Thoughts
A sincere thank you to Bisnow, our fellow panellists, and everyone we had the opportunity to meet throughout the event.
The conversations were insightful, the perspectives were diverse, and the appetite for meaningful innovation across the built environment remains incredibly strong.