Artificial intelligence and machine learning in facilities management?

Technological and scientific advancement has brought some hugely exciting changes to our lives and this looks set to continue. 

Perhaps one of the most potentially exciting is that of artificial intelligence (AI) or machine learning (ML), though as you will see below this is actually two things!

Like the Internet of Things (IoT), AI/ML has for many appeared to be more hype than substance. Many companies have been born promising to change the world through artificial intelligence and forecasts about when we will achieve human level ‘general intelligence’ vary greatly, even amongst the most knowledgeable on this subject. 

Facilities management and property maintenance, like most industries has been caught up in this excitement. 

But how can AI/ML actually be used in facilities management?

The difference between artificial intelligence and machine learning

Whilst often used interchangeably, artificial intelligence and machine learning are actually different things. To be specific, machine learning is a subset of artificial intelligence (or a means by which artificial intelligence can be achieved). 

There are a number of ways in which artificial intelligence can be considered to have been achieved. Perhaps the most famous of these is the Turing Test in which a person for example has a conversation with another entity that it believes to be human but is actually a machine designed to produce human-like responses. For the Turing test to be passed, the machine must be considered indistinguishable from a human. 

In deciding what responses to provide, the machine will (likely) undertake a process called natural language processing, whereby it assesses the meaning of what has been ‘said’ to it and provides an appropriate response. Natural language processing merely refers to the area of AI associated with language. 

Machine learning relates to how decisions are made. In order to display ‘intelligence’, the machine must make decisions. These decisions are made in the context of an expected response (e.g. continued conversation) which themselves have been determined by looking at the previous results from the same (or similar, or comparable) decisions. 

From this you can see that for artificial intelligence to be demonstrated, there must be data that is available to be analysed. How this is done, and continues to be updated in light of additional data is where machine learning comes into its own. It is, perhaps unsurprisingly, how the machine learns. 

Potential applications of AI/ML in FM

Now that we have provided a working definition of artificial intelligence and machine learning, we can have a look at where it already (and where it will) have an impact on the world of facilities management. 

1. Predictive maintenance

As you can likely tell from the above, artificial intelligence relies on data to make decisions on actions to be taken. By monitoring, for example on a continuous basis via sensors, plant and equipment data can be obtained with regard to equipment functioning. When to maintain, repair or replace can then be a decision that benefits from AI/ML.

As buildings such as The Edge in Amsterdam deploy sensors at large scale (more than 30,000 in this case), huge amounts of data are generated related to how the occupants interact with the space and the building. Using this information can prompt building owners or managers to reconfigure spaces to suit the needs of inhabitants. 

In addition to space optimisation, where for example parts of the building can be closed and employees directed toward open desks in particular areas in response to occupancy levels, environmental factors such as air quality or temperature can be changed based on data. For many companies, optimising the work environment is a means of increasing productivity and AI/ML can already be used in this area. We will no doubt see continuing advancement in this area. 

One of the most promising applications of AI/ML in the world of facilities management is that of energy consumption. This has both cost and environmental benefits and is already in use by companies such as Google, who for example were able to reduce the energy consumption in their data centres by 40% by utilising AI/ML. For example by building a model of occupancy, temperature and  coupling it with weather forecasts, a building’s energy management system can decide to reduce heating or cooling. It is the improvement of this model that capitalises on AI/ML. 

More relevant with regard to integrated workplace management systems (IWMS) than computerised maintenance management systems (CMMS), AI/ML has tremendous scope to reduce the time and resources required to handle things such as loan agreements, leases and contracts. Pertinent and timely information can be surfaced for facilities managers by analysing these lengthy and complicated documents. 

Summary

Facilities management, like so many other areas of the business world is likely to be hugely changed by artificial intelligence and machine learning. 

At present, the applications are somewhat limited but targeted at tasks that can have a demonstrable return on investment. This is for the best. 

AI/ML is not something that can just be unleashed on a role or function and will eventually supersede the need for the incumbent resource. Facilities management is a complex industry which requires continual tradeoff and assessment. This makes it difficult to pass responsibilities over to a machine, but on the plus side it is also an industry that produces huge amounts of data that can be used to achieve targeted objectives. 

Welcome to the brave new world.