In the midst of a smart building revolution, smart, or autonomous, buildings are the end goal. The creation of smart buildings and cities which maximise energy efficiency is essential in the current global climate. To dial down into some more depth on this topic, I recently invited experts in the space, Katie Whipp, Will Brouwer and Abel Samaniego to discuss how to achieve fully autonomous buildings using artificial intelligence (AI).
During this interactive webinar, as part of our CM Industrial series discussing The Future of Smart Cities, we touched on how 99% of the world’s buildings are actually lacking the Internet of Things (IoT) infrastructure needed to create smart buildings. Even within the climate-conscious UK, only 34% of buildings are using smart technology. However, widespread incremental AI utilisation means that fully autonomous buildings, and greater energy efficiency, have real necessary potential.
When it comes to effective data collection Katie Whipp, Head of UK at Deepki, commented:
It’s really important to recognise that bad data leads to bad AI technology… But I think equally, too much data also lends itself to AI struggling. That’s because it’s not able to create the patterns, the structures it needs to learn the practices that we want AI to put in place and automate.”
Therefore, the key to creating fully autonomous buildings with AI is access to “good” data. She added that the building needs the ‘goldilocks’ amount – too much and useful overarching patterns will not be generated. To get the right quality and amount of data, identification of data sources and their health (in terms of reliability and relevancy) is essential. To improve data quality, she recommended avoiding manual data collection and extracting it as close as possible to the source, as the former will inherently contain errors.
Abel Samaniego, CEO & Founder at DABBEL Automation, agreed, adding that:
99% of the buildings in the world don’t have a highly smart building IOT infrastructure. So, how can we solve it? You need data. You need to be tagging. You extract the data and then in many cases you will train your algorithms based on what you read from the building. You won’t probably have enough to provide the self-adaptation of the learning.”
Correcting the lack of IoT infrastructure in the world’s buildings will firstly require the collection of data from buildings to train these algorithms. He added, however, that this would not drive learning alone, meaning other components are necessary.
...Facility managers are always the busiest person that I ever see in the world of property management."
However, she is passionate that they are at the forefront of the smart technology and data revolution. These assets have an advantage for everyone, making them a shared responsibility rather than that of one team.
Will Brouwer, Senior Product Manager at WiredScore, noted that using AI at the facility manager level first and foremost improves building sustainability.
...Predictive maintenance is another good example, which provides increased equipment uptime and reduces costs. This is where an AI system or machine intelligence predicts when a machine is likely to fault…
It then prompts facilities management to investigate and maintain that system before it fails.”
Not only does predictive maintenance predict when facility management teams should fix faults to increase equipment uptime, and reduce as a result, the insights and building tuning recommendations provided by AI help reduce energy consumption.
Brouwer also explained how cost efficiencies, future-proofing and the inspirational experience provided by the building are all improved when facility managers integrate AI into their work. Advocating the use of facility managers, he explained how the amount of data collected by an AI proportionally increase the team’s value to other building users and the environment.
A recent IBM Institute for Business Value study supported these findings, reporting that 76% of COOs felt increasing facility management automation increases operational efficiency. Additionally, 99% of real estate respondents believe smart technology is essential for their roles. It seems facility managers (however busy they are), and fully autonomous buildings, do need AI.
With data being such an integral part of AI, data loss or inaccuracy are some of the worries with using AI for autonomous building. Katie’s company, Deepki, uses a short-term solution where AI can backfill gaps using historic data, meaning data blips are not a problem.
However, a more continual and emerging barrier to autonomous buildings using AI, especially as increasing numbers of them are connecting to the internet, is maintaining cybersecurity. A topic my colleague, Bill, touches on in much more detail in a follow up webinar.
With only 1% of the world’s buildings having the IoT infrastructure required for autonomy, the implementation of fully autonomous buildings must be completed in phases. However, autonomy encompasses a building’s entire lifecycle, from design and assembly to the experience of and interaction with its users.
Where the creation of fully autonomous buildings with AI is the end goal. The first, and necessary, focus is the operational alteration of existing buildings to reduce the emissions which are killing our planet.
Fancy chatting about smart buildings, the future of proptech, or the wider implementation of AI across the industry? Drop me a message at firstname.lastname@example.org or connect with me on LinkedIn.
Where the creation of fully autonomous buildings with AI is the end goal, how do we get there? I invited three esteemed industry leaders to discuss just that. Click to read their insight.