We are entering an era in which artificial intelligence (AI) will play a profound role in facility management processes.
Survey – Artificial Intelligence in Facility Management Services
Thank you for your interest in taking our survey about artificial intelligence in facility management processes. The survey is now closed, and we are using the input to create “The Definitive Guide to AI in FM”. If you would be interested in receiving the survey results and this guide when it becomes available, please fill out this form.
Trying to define what AI actually is has resulted in many different descriptions. For example, the word “robot” should not always be taken literally: robots can be made of software only. In fact, as “artificial intelligence” tools become common place, people stop thinking of them as AI. Take autofill as an example: people have become so used to Google accurately guessing what they want to search for after typing a few words or letters, that it’s no longer a “wow” factor – it’s an expectation.
I believe the term “artificial intelligence” is sometimes used incorrectly, similar to the word “smart” as I have written about before. I prefer the term “machine learning” when discussing the role of AI in facility management.
Types of artificial intelligence
In a survey my colleagues are currently conducting, they are asking facility management service providers to share information about how they are utilizing the following areas of artificial intelligence:
- Natural language processing. For example, reserving a room or ordering catering by talking to a digital personal assistant.
- Computer vision. For example, using security cameras that automatically identify people and create an alert after the building has been closed.
- Planning and scheduling. For example, having an advanced algorithm calculate the best scheduling of skilled workers to the right tasks based on optimized travel routes.
- Robotics. For example, using drones to perform inspections in dangerous locations.
- Decision support. For example, using big data analytics to predict and respond to potential asset failures and improve the profitability of client contracts.
- Machine learning. For example, service ordering systems that learn employee preferences and start to make recommendations or automatic adjustments based on that information.
- Process automation. For example, having a platform that uses contract data and customer preferences to automate the process of being compliant with customer contracts.
Computer vision is a particularly interesting category, especially when considering the role it can have at facility management service providers. Service providers need staff that can be effective and profitable quickly. Depending on the service provider, human resources can account for up to 80% of costs.
This is why it is interesting to explore artificial intelligence. Computer vision takes a quality that was previously exclusive to humans – both seeing and understanding what was seen – and expands this skill to software. An interesting example, includes publications about computer vision and the sensor systems used to interpret the data imply that cameras are becoming smart. These cameras not only stream images, they are able to augment those video streams with streams of data describing the things that are seen. The technology applied to allow cameras to learn and interpret the images is called “Deep Neural Network” or DNN.
The learning capability of these cameras is based on neural network technology and is in some cases being hard-wired, i.e., implemented on a silicon chip, which is mounted in the device itself.
AI augmented technology can solve several staffing challenges for service providers. For example, one area of current research is autonomous driving. Autonomous vehicles require high amounts of sensory inputs to assess conditions of their environment. This type of vehicle is already being deployed at warehouses, automating part picking and placement without human intervention.
Cameras are a common sensor in buildings, as well. Applying smart cameras around buildings will create new use-case potential for real estate, facilities, and security management. Learning capabilities can be applied to facial recognition, which could allow for refined access management but also to ‘count’ the number of people in areas of the building or even to identify events happening. Global research advisory firm, Gartner, reported that systems are now more reliable in recognizing individuals by the shape of their faces than humans are. This means these computers can help reduce security staff working overtime.
We have seen use cases where smart cameras are deployed to identify “events” like the illegal dropping of waste and the use of elevators in undesirable (unauthorized) ways. This allows for immediate response on this type of event, preventing damages or costs.
However, the biggest caution area to explore and to take seriously when implementing systems like these is privacy. In its 2019 ITXpo publications, Gartner research points at the fact that among users of technology, privacy ranks higher in importance than convenience.
The principles of smart technology in cameras can, and probably will, be applied to other devices and use-cases as well. Is this one of the areas that you and your organization are already exploring in AI? Do you expect to start using it in the next two years? Let us know in our survey about AI in FM processes.
Global Product Strategy & Innovation Director