Chatbots Heart of Smart Facilities Management

Why Chatbots should be at the Heart of Smart Facilities Management

Posted on Posted in Facilities Management

This article first appeared in the International Facility Management Association (Singapore) May 2019 Newsletter

In recent years, terms such as Smart Cities, Smart Buildings and Smart Facilities Management have become popular. These terms evoke visions of artificial intelligence, data analytics, physical robots, ubiquitous sensors, video analytics, automated energy optimization and fully automated workflows, amongst others. On the ground the reality is often different, with paper checklists, excel files and manual processes still prevalent.

How do we get from today’s ground reality to the vision of Smart Facilities Management? This article contends that chatbots should be at the heart, and indeed the starting point, of efforts towards this vision.

Chatbots are the human – systems interface

Chatbots use artificial intelligence powered natural language processing to parse human language into data fields which can be processed by digital systems. It is useful to think of chatbots as simply a way for humans to “talk” to a system in a natural yet automated way.

Typically, chatbots are deployed on common messaging platforms, such as Facebook Messenger or Telegram. Whatsapp is expected to also allow chatbots to be deployed in the near future. Just as many of us now prefer to message our friends rather than make a call; communication with the facilities management system, such as reporting a fault, will also move towards messaging rather than voice calls.

Happily, this user preference also brings benefits for smart facilities management. Since messaging is digital in nature, the fault reporting process flow becomes digital from end-to-end; allowing data to be seamlessly collected, response times to be accurately measured and other digital features, such as language translation, to be easily implemented.

Who are we serving?

Visualizing chatbots as the human-systems interface brings into focus its importance. After all, the ultimate goal of smart facilities management systems should be to benefit the people using the building. Whether it is preventive maintenance to make sure chiller systems do not fail, or repairing room lighting, the goal is to allow the people in the building to live or work productively and comfortably, and also reduce the cost of doing so.

Before the advent of chatbots and messaging systems, there was an analogue gap between humans and the facilities management system. There was no good way for facility managers to digitally communicate directly with the people using the building. The people who should be the ultimate beneficiaries of the facilities management systems were ironically not part of its digital eco-system.

Humans are sensors too

While installing a range of Internet of Things (IOT) sensors in buildings will undoubtedly be beneficial for automation, we should not neglect that the most important system feedback should be from the people using the building.

Artificial intelligence tries to mimic human intelligence, but is still far from the real thing. For example, humans can exercise discretion on whether a bag is actually suspicious and whether the bag should be reported. An IOT device may be able to report the temperature in a room, but it cannot detect the climate preference of the humans in the room.

Data collected from IOT devices installed in smart buildings should be augmented by conversational data recorded from human feedback via chatbots in order to arrive at better automated outcomes to benefit the people using the building.

Data Collection before Artificial Intelligence

Data collection is the starting point in the development of any artificial intelligence system. In order to arrive at the vision of the smart building that this article started with, it is necessary to go through the phase of data collection.

Non-building specific data can be very helpful. For example, a lift manufacturer may be able to aggregate data across all its lifts and implement artificial intelligence to try to predict when the lift may fail.

But since each building is unique, non-building specific data should be augmented by building-specific data. For example, the building environment in which the aforementioned lift was installed, its usage characteristics and preventive maintenance schedules may affect the prediction. Therefore, it is important for each building to collect its own data through a data hub.

This is a role that a chatbot-powered facilities management system is well-suited to perform. Starting with the most important data – conversational data from people using the building – such a chatbot-powered facilities management system can accept data from other sensors and IOT devices, in order to form the basis for building-specific artificial intelligence to be applied.

Chatbots Offer a Route toward Smart Facilities Management

As the facilities management industry in Singapore moves inexorably towards the vision of Smart Facilities Management and Smart Buildings as building blocks for Smart Cities; it is important to consider what steps to take to arrive at this vision.

Given the twin forces of shifting user preference towards messaging and the maturity of natural language processing in chatbots, implementing a chatbot for facilities management is a cost effective and practical first step towards this vision. Since the chatbot is a software implementation, it does not involve large equipment purchases or installation costs. Productivity benefits for the facility management and convenience benefits for the people using the buildings can be quickly felt.

As the Chinese saying goes, “A journey of a thousand miles begins with a single step”.

This article has been contributed by Patrick Sim, Co-founder and Director of RobustTechHouse, which has developed FacilityBot. More information about FacilityBot is available at

Photo by from Pexels

Also published on Medium.

Leave a Reply

Your email address will not be published. Required fields are marked *