From Fear to Curiosity: How AI Is Reshaping the Built Environment

What if the story of artificial intelligence in the built environment is not one of human replacement, but of human empowerment? 

This is the thought-provoking question at the heart of a recent conversation with Maryrose Lyons, founder of the AI Institute. With a background blending digital marketing, user experience, and cyber psychology, Maryrose offers a refreshingly human-centric perspective on the AI revolution.

She argues that the narrative is shifting from fear into an era of curiosity, and more importantly, that this shift is fundamentally changing how professionals work and what they value.

The Quiet Revolution: How Language Around AI Has Changed

Not so long ago, the discourse surrounding AI was dominated by anxiety. As Maryrose notes, the language was steeped in fear: fear of job losses, fear of being replaced, and for some, existential dread. 

This gradually gave way to compliance concerns, with data security becoming the primary worry. Then came overwhelm, as the pace of technological change left people struggling to keep pace.

But now, a new word is emerging: curiosity. 

"This is the era where if you have a curious mind, you are so going to succeed in this new world," Maryrose asserts. 

She has been tracking this linguistic shift with her teams, observing how the language people use when learning AI has evolved. 

The progression from fear to compliance to overwhelm to curiosity reflects genuine maturation, of people beginning to understand that AI is not a threat to be managed but an opportunity to be embraced.

The curious mind, paired with action, creates the conditions for genuine transformation. "You're the best people because you do incredible stuff," she notes, describing how individuals who approach AI with genuine inquisitiveness consistently find novel applications that go far beyond what any training programme could prescribe.

Empowering Humans, Not Replacing Them

At the core of Maryrose's philosophy is a principle that stands in stark contrast to much of the AI narrative: "do more with same."

This is not about cutting costs through redundancies. Rather, it is about empowering individuals to achieve more meaningful work with their existing resources and, crucially, to reclaim their time.

How This Works in Practice

When Maryrose works with organisations, she follows a deliberate process. 

• First, identify a time-consuming task. 

• Second, learn how to use AI to accomplish it more efficiently. 

• Third, and crucially, ask: "What can you do with the time you've just saved?"

  
  

Consider a concrete example: creating a LinkedIn carousel. This once involved moving between Canva, copying and pasting content, resizing images, and adjusting layouts. With AI, Maryrose can now generate assets, train the AI on her organisation's design style, and produce a polished carousel in a fraction of the time. 

She only needed to make changes to one slide out of ten.

But here is where the philosophy becomes powerful. The time saved is not reinvested into doing more of the same work. Instead, it becomes an opportunity for deep thinking, learning, meaningful collaboration, or simply taking a break. 

This requires a fundamental shift in mindset at the leadership level, one that goes against decades of productivity-focused thinking.

Beyond Training: The Use Case Workshop Approach

Maryrose discovered that approximately 80 per cent of organisational problems can be solved by people trained in AI fundamentals. The remaining 20 per cent are different: deep, systemic challenges that have evolved over time, the "way we've always done things."

Voice-Driven Surveys and Bespoke Learning

To understand these deeper challenges, her team developed voice-driven surveys. Rather than tick boxes in a Microsoft form, professionals speak about their challenges, frustrations, and ideas for automation. 

"We get these much deeper, richer answers," Maryrose explains. People open up about areas where they are stuck, about creative ideas they have been mulling over.

Once training begins, participants receive homework assignments applying what they have learnt directly to their own work. On Friday mornings, they gather for show-and-tell sessions where they demonstrate what they have accomplished. These sessions are where the real magic happens. 

Professionals share solutions, and others see possibilities they had not considered. "Imagine if we could do it this way," someone will suggest, and course instructors furiously take notes.

This approach recognises something fundamental: most people genuinely want to do their jobs well. When given the tools and permission to experiment, they are remarkably creative. 

"You very rarely get like two people doing the same thing," Maryrose observes. "It's almost like limitless the amount of solutions you can do."

The Fluid Future of Work

We are witnessing a fundamental transformation in how work itself is structured. 

The traditional model, where processes are defined from above and individuals follow them precisely, is giving way to something more fluid and adaptive.

In this new paradigm, individuals become the "CEOs of their roles." Rather than asking "What process should I follow?", the question becomes "What is the outcome I need to achieve, and what is the best way to get there?" 

Process becomes a tool in service of outcomes, not an end in itself. 

A curious professional, armed with AI tools and a clear understanding of their objectives, can think critically about the best approach, gather necessary knowledge, craft a well-designed prompt, check their work with human oversight, and deliver the outcome.

This shift has profound implications for software companies. Traditional SaaS platforms assume users will adapt themselves to fit prescribed workflows. But as AI becomes more capable and flexible, this model is breaking down. 

Why should you restructure your data to fit a SaaS tool when an AI system could simply understand your data as it exists?

The Real AI Agent: Separating Hype from Reality

One of the most misunderstood concepts in the current AI landscape is the notion of an "agent." Many SaaS companies claim to have agents, but Maryrose is quick to point out that much of what is marketed as agents are actually virtual assistants.

A true AI agent is characterised by autonomy. You give it a task, and it goes off and does it itself without requiring manual data movement between systems. A real agent can monitor a spreadsheet, detect changes, pass information to another system, process it, and notify you when done.

Consider a practical example: an agent that monitors government tender sites for opportunities matching your criteria, downloads relevant documents, pre-fills response forms using predefined templates, and passes the 80 per cent completed form to a human who can add expertise and submit it. 

This is genuine autonomy in service of human work.

The problem is that many vendors use the word "agent" to describe sophisticated chatbots or virtual assistants. These systems answer questions but lack the autonomy that defines a true agent. 

Interestingly, Microsoft was probably using the word "agents" knowing true agents were coming down their roadmap. The agents now released through Copilot Studio are indeed genuine agents, not the virtual assistants users create through standard chat.

The Built Environment Advantage

Maryrose's recent report on AI adoption in the built environment reveals surprising findings. Architects are leading adoption, knocking it out of the park with their use of AI tools. 

However, the most striking finding concerns company size: mid-sized companies are winning the AI race, not the largest enterprises or smallest firms.

Very large organisations have been hamstrung by licensing constraints and complex procurement processes. Very small firms, overwhelmed by running a business, have struggled to find time to experiment. 

But mid-sized companies, with enough resources to invest but nimble enough to move quickly, are thriving.

One of the most striking findings is the extent of shadow AI. Professionals across the built environment are using AI tools on their own devices, often without formal organisational approval. 

This creates security risks and means organisations are missing opportunities to learn from and scale successful applications.

Perhaps most tellingly, the research identified 74 different AI applications being used, far beyond the large language models dominating headlines. This speaks to the creativity of professionals finding novel applications to solve their specific challenges.

Looking Forward

As remarkable as progress has been, Maryrose believes the most important development on the horizon is one few people are talking about: spatial AI. Fei-Fei Li, the godmother of AI, has been working on spatial AI for about a year. More recently, Yann LeCun left Meta to establish a spatial AI startup.

For the built environment, this is particularly significant. 

Architects and engineers work in three-dimensional space. Spatial AI could revolutionise design workflows, allowing professionals to interact with digital models in more intuitive and powerful ways.

"2026 is the year you can use the AI we've got to solve all those businessy admin boring problems because spatial AI is coming and that is going to fundamentally change everything," Maryrose advises. 

Use the current generation of AI tools to automate routine work and free up your time, because when spatial AI arrives, you will want to be ready to harness it for the work that truly matters.

Why Is The Full Conversation Worth Your Time?

This conversation with Maryrose Lyons barely scratches the surface of the insights she shares about AI adoption in the built environment. 

In the full podcast episode, you will hear much more about the specific strategies she uses to help organisations navigate the cultural and technical challenges of AI implementation. 

You will discover the details of her AI adoption report and what the data reveals about which professionals and organisations are leading the way. You will also explore the nuances of true AI agents, the promise of spatial AI, and the role of cyber psychology in managing change. 

Most importantly, you will hear from someone who genuinely believes in the potential of AI to make our working lives better, not worse. 

To listen to the full conversation, visit the Project Flux podcast.

All content reflects our personal views and is not intended as professional advice or to represent any organisation.