Kardham Digital launching KD Forecast, an algorithm based on AI and machine learning capable of predicting, 30 days in advance and with over 90% accuracy, the occupancy of a commercial building. Operators of business premises will be able to rely on this tool to offer users an easier experience of their workspaces, while optimising the management of these spaces in a context of rationalising property costs and reducing their carbon footprint.

A paradigm shift from m² to flow analysis

From an organisation of work in a dedicated office where each person likely to come into a workspace had a permanent place, teleworking and flex-office, two characteristics of the post-Covid years, have brought companies face to face with a new challenge: that of having to manage an organisation where the flow of people using these spaces varies every day. And in a new paradigm: whether we’re talking about energy consumption or carbon footprint, the intrinsic metrics of a commercial building are relatively constant during working hours, while occupancy by flow shifts the unit of account of that same building from its “capacity” to its “use”. What characterises a building from now on is more than ever its users, their arrival and their use.

Predicting occupancy: a central element of optimisation strategies

Highlighting the number and activity of people present on site as the main variables characterising the building is a valuable lesson: to be able to optimise the m2 and energy consumption of their workspaces, companies will need to have a detailed analysis of their occupancy. This is precisely what KD Forecast offers. In addition to using data relating to surface areas, floor plans, capacity and building management data, KD Forecast uses tens of thousands of building occupancy data points and more than 30 peripheral parameters to anticipate the number of users who will occupy these spaces with the greatest possible accuracy and time projection. This data includes :

• Traffic data: occupancy of workstations, collaborative spaces, number of people per zone, etc.
• Access data: many commercial buildings are equipped with access control systems that provide reliable presence data.
• HR, IT and FM data, the heart of the system data relating to users: HR repository, space planning, O365 access, etc.
• Booking data: meeting rooms, intended attendance, catering, etc.

Please note that no personal data is used in KD Forecast.

KD Forecast is the result of over a year’s research by Kardham Digital’s R&D Data Department. Thanks to the Gradient Boosting method, the tool outperforms calculations based on simple observation of the past (time series method) by 60%, achieving prediction accuracy of over 90% on average.

Its mathematical modelling is at the crossroads of a number of issues:

• Identifying the data to be used: the possibility of defining it, collecting/producing it, processing it, etc.
• Choosing the actions to be taken on the data: defining the rationalisation of the data between them, the applicable treatments, the creation of value in the interactions, etc.
• The use of reprocessed data: defining the choice of decision-support dashboards, contextual information systems, the depth of time required for anticipation, linking building automation systems to actual usage, etc.

As well as improving the economic and environmental optimisation of workspaces, KD Forecast offers other advantages for workspace managers: adapting their site to the expected traffic in terms of both services and the spaces made available, anticipating periods of potential under/over-occupation to better manage arrivals when necessary, taking into account the evolution of the site when the company grows or reorganises, etc. For users, KD Forecast promises to be a tool that provides a smoother, simpler experience of workspaces, for example by eliminating the need for a daily attendance declaration that is both uninspiring and unintuitive.

“All these issues around AI as an innovation for better building management to optimise energy consumption and/or rationalise operation are not new at all, they have been around for over 10 years (although they are still not widely deployed). What is new is knowing what real leverage we have today if we want to go even further in optimising these buildings. To do this, we need to look at what they are used for: for us, the building is a service centre, that is its purpose. To optimise it, these users must be the starting point for thinking about and developing any innovation. The most important leverage effect in terms of rationalising m² and reducing the carbon footprint of a site is the occupation of that site. That’s what we wanted to do with KD Forecast: consider that the individual should now be the pivot of building data”, Jérôme Hérard, Director of Digital Workplace Solutions at Kardham Digital.

“The launch of KD Forecast is further proof of our conviction that digital technology is a key area for innovation, enabling the industry to move towards greater economic, environmental and social performance. Our vision is to make buildings increasingly autonomous, without human intervention in their day-to-day management, while at the same time enhancing the user experience. To achieve this objective, the building information system needs to be able to predict what is going to happen in the building so that it can better anticipate the services that need to be provided to the occupant. Today, everything in our industry is changing very fast: the operators of professional spaces, the users, the professions, the ways in which people work and the uses to which they put their time. These are all needs and interests that digital technology can help bring together. This is what led to the creation of Kardham Digital more than four years ago, and the starting point for each of the projects that our team of 80 technicians and engineers designs for our customers”, Pascal Zératès, Managing Director of Kardham Digital.