Built environment technology in 2025: Between LLMs and PDFs

Tuesday 04 March 2025

Founder
Occasio Insights Ltd

Built environment technology in 2025: Between LLMs and PDFs

Camilla Siggaard Andersen, Chair of the Built Environmental Technology Expert Panel, reflects on their first meeting of the year, reviewing past achievements and key themes for the year ahead.

It’s been more than three years since the first meeting of the NLA’s Built Environment Technology Expert Panel. Since then, over 30 individuals have contributed their time and expertise to advise on the use of technology in shaping London’s built environment. Every cycle has yielded valuable insights, documented in whitepapers, blog posts, and engaging panel discussions. If you’re curious to revisit our past discussions, explore the links at the end of this post.

On Monday 10th February 2025, we kicked off the fourth cycle, welcoming a dynamic mix of new and returning members. Our current group brings together diverse perspectives, from strategic public sector stakeholders and ambitious solution providers to tech-savvy practitioners and industry innovators. The list of all panellists can be viewed here.

We began our meeting by reflecting on the panel’s accomplishments over the past three years. The landscape has certainly shifted. The mainstreaming of AI, particularly Large Language Models, has profoundly impacted how we work and live, prompting professionals, including architects, engineers, developers, contractors, and service providers, to re-evaluate their roles in the future economy.

However, some challenges persist. The panel continues to express frustration with the predominantly analogue nature of the planning process, which often relies on static PDFs rather than dynamic, collaborative virtual environments.

It’s remarkable that while we contemplate the potential for digitally designed and managed cities, we still struggle with the limitations on PDF file sizes imposed by local planning authorities.

At the heart of our discussions about digital technology lies the crucial issue of data – specifically, the quality and quantity of data extracted from and deployed within the built environment. Several key challenges continually surface:

The price of information: Gathering, processing, and utilising data demands significant investments of time, capital, and natural resources. Collecting everything is neither economically sensible nor environmentally sustainable. Yet, we haven't reached a consensus on which data points are truly essential and which we can afford to ignore. Forecasting future data requirements adds another layer of complexity.
The ownership question: Data, like any valuable asset, requires stewardship and maintenance. However, concentrated ownership risks exacerbating inequality and creating uncompetitive markets. This concern was articulated in a 2020 open letter to Autodesk, signed by British and international architects, which highlighted the firm's growing influence on architectural software development. With the rise of cloud-based and subscription models, this issue has only intensified.
The quality imperative: The adage “garbage in, garbage out” rings true for urban data models and property technology. Ultimately, a lack of consistent data standards within the built environment professions hinders the development of effective digital tools. This problem is compounded by a scarcity of reliable data sources to fuel these models and ensure consistent, accurate outputs. You can’t make sound decisions with flawed information. These challenges are inextricably linked to the questions of prioritisation (point 1) and ownership (point 2).

Data management and utilisation tend to be most effective when there's a clear business case and demonstrable benefits for providers, operators, and end-users alike. The integration of public transit data with Google Maps, facilitated by the General Transit Feed Specification (GTFS), provides a compelling example. By adhering to GTFS, local transit agencies enable Google Maps to seamlessly guide their customers, eliminating the need for each agency to develop its own navigation app. During the pandemic, Google even shared anonymised data back with local authorities, providing valuable insights into changing travel patterns. For a time, the cycle of data sharing and benefit was complete.

At the end of the discussion, the panel identified three key themes to focus on this year:

The future of the profession: How will architects work in 2040, as they interact with developers, contractors, communities, and increasingly sophisticated AI? We will explore three extreme scenarios to illuminate potential future pathways. Key areas of focus include skills, training, automation, job creation, and the development of new service offerings.
Laying the foundations for built environment data.: What foundational elements must the public sector establish to ensure the safe, useful, equitable, and sustainable long-term production, storage, and utilisation of built environment data? We will examine three best-practice examples and propose three priority initiatives. Key themes include infrastructure, data standards, urban resilience, regulation, and prosperity.
Bridging the data gap in practice: Why is so much valuable built environment data, generated by architectural practices, service providers, developers, and local authorities, ultimately lost or underutilised? We will map the current data value chain and envision an ideal future state. Key considerations include data quality, data availability, data exchange, responsibility, and commercialisation.

We will reconvene in June to delve deeper into these crucial topics. If you have relevant case studies or perspectives to contribute, we encourage you to contact us.