Beyond ChatGPT: AI for sustainable construction with Contilio
As we reach the one-year anniversary of ChatGPT, we reflect on the impact it has had on the built environment and wider society. On one hand, it has revived the conversation around ethics, misuse of buzzwords to attract capital, and technology’s potential to supplant human industry. On the other hand, reports from real estate innovation managers indicate that ChatGPT has been a catalyst for change within their organisations. What is often missing from the 12-month discussion around ChatGPT and generative AI more broadly is the long history of AI in the built environment and use cases beyond chatbots and image generators. As experienced AI entrepreneurs, Contilio’s Georg Glatz, Zara Riahi and the wider team have shared their thoughts with us below.
What is Contilio and how did it come about?
Contilio is a generative 3D AI analytics and insights platform solving the most critical problem for the $12 trillion construction industry: that is the complete lack of timely, connected insights on core construction processes, which for a $1 billion dollar portfolio results in up to $300 million of rework costs, significant delays and associated costs (construction projects are on average 30% late), and 1 million tonnes of avoidable waste and CO2 emissions.
Contilio solves this problem by providing for the first time connected, scalable, and powerful insights on Quantity Tracking, Progress, Quality Verification, and Risks in fully automated manner and from one platform and generating a Cognitive Digital Twin previously missing from construction and operation phases. This allows real estate developers, general contractors and asset owners globally to save millions in delay, rework and maintenance cost, wasted manhours, and most critically CO2 emissions.
By bringing together a founding team with experience applying 3D AI to real world problems in healthcare, finance and other industries, we quickly homed in on this $5 trillion issue for the world’s largest industry which is of huge importance for society and for each and every individual. Beyond solving its core problems, our platform also creates new opportunities, including faster time to payments, green financing, and lower dispute and operating costs, making everybody better off.
How does Contilio differ to ChatGPT and other generative AI tools out there?
The generative pre-trained transformer technology behind ChatGPT has achieved an impressive maturity but may be used to create value for use cases where there is a lot of siloed knowledge, without the need for being highly accurate. These include copy writing, replacing Google searches, or creating first-level customer service agents. What tends to be ignored is that ChatGPT and its underlying technology give you a likely answer at best, that is based on a large and costly volume of training data. There is no guarantee that the information is correct, something that is a key requirement in most industrial applications. Also, GPT’s models focus on language models and some 2D image and video applications, but 3D vision applications are very limited. This is where Contilio’s product differs.
Contilio uses 3D AI to generate the missing insights from scratch and to bridge the gap between the built reality and the digital representation, generating for the first time a cognitive digital twin that is machine-readable, accurate and always up to date. But we don’t stop there; we also generate augmented building information models (3D design models with millions of complicated elements) and schedules (project plans with thousands of entries), enabling our customers to access accurate insights, and a traceable, data-driven single source of truth.
This is done by using cutting-edge 3D AI and computer vision techniques (patent pending) to make sense of the built environment in 3D and assign meaning to every point of a lidar scan which is otherwise meaningless and purely visual. It then connects this intelligent reality to the BIM and schedule, and automatically calculates progress, quality, and quantities along a hierarchy starting from elements to groups, trades, functions, floors, and up to projects and portfolios. Unlike ChatGPT, we have introduced a constrained & retraceable generative approach that ensures we are building on reality, adhere to the laws of physics and construction requirements, and can trace how insights are generated. That is a key feature of our technology, crucially important for any reliable AI solution with industrial industries, including in manufacturing, logistics or aerospace.
AI ethics has been a hot button topic for a long time, recently popularised by the likes of Elon Musk. How does this apply to Contilio and other AI-enabled construction tech?
As with every new and fast advancing technology, AI brings about opportunities and risks that need thoughtful management. Data privacy, the individual’s right to be protected from intrusive analysis or targeting are key concerns, as are clear rules around data ownership and usage for learning purposes.
The Contilio team combines decades of experience handling sensitive data in a secure manner thanks to our years building medical vision software and in finance. Based on that experience, we made data privacy and information security a key pillar. We clearly define data ownership with our customers and automatically remove non-construction elements (such as humans) from customers’ data.
One concern that is often raised is that “automation” leads to worker redundancies. Our return-on-investment (ROI) analytics show that the time our customers’ teams save doing manual, mundane tasks, can be spend on the value-adding tasks that they have been hired to do–in other words, (i) using the insights to make much better decisions in a timely manner; (ii) use their time to resolve issues and risks; and (iii) to devise better plans and forecasts. We strongly believe that Contilio and similar technologies will help our industry attract talent and gradually close the widening skilled worker gap that it suffers from.
What environmental impact could AI have in the built environment?
We have not even scratched the surface of what 3D AI can do in terms of CO2 reduction and removal for the built environment, which contributes roughly 43% of total greenhouse gas emissions globally. While improving design, energy efficiency and the use of green materials are great initiatives, the elephant in the room remains that up to 30% of construction expenditure is wasted, representing trillions of dollars of value on a global scale.
Currently, the industry works with backward looking metrics, months old, siloed, and sometimes inaccurate information, providing visibility into less than 10–20% of what is happening on-site. This results in 8 out of 10 construction issues not being found, or found very late, leading to costly rework, and huge amounts of waste. These mistakes, and a digital representation that does not match the reality, translate into further inefficiencies during the asset’s operation over many years. Contilio solves this, providing the much-needed visibility and key performance indicators (KPIs) in a manufacturing style, and fully automated manner, making it possible to save millions in costs and thousands of tonnes of CO2.
We and our customers are also hugely excited about the upcoming release of our AI CO2 tracker, building on the same platform and data. It enables our customers to accurately track and remove CO2 from their construction portfolios and also operating assets, reducing CO2 and related taxes and fines, accessing green financing opportunities, especially at times that legacy loans are maturing, and companies need to protect their asset value and revenue streams.
You can learn more about Contilio via their website: http://contilio.com
This article forms part of Pi Labs’ upcoming paper Sustainably intelligent: AI for a greener built world. Register your email address to be among the first to receive it to your inbox upon public release.
