Espoo tests artificial intelligence in property management
The City of Espoo is responsible for the servicing of a large number of properties. The city was debating the possibility to use artificial intelligence for data transfer in the planning of work and in the building activities in the city.
What kind of problem is the project trying to solve?
The City of Espoo has hundreds of thousands of sites it maintains and is planning to build on. Maintenance applies to the entire infrastructure. Until now, data has been entered into the systems manually, which is slow. As a result, systems contain a lot of mistakes, the information in them is not up-to-date and mistakes have been made in interpreting the entries made by different people.
The Trello project management programme contains a task queue for the city's building activities. For example, plans typically go through 9 stages in which they are drawn and validated several times. Because of the complicated system, time and resources are wasted during maintenance measures and planning.
Objective: Automatic transfer of data to the register
The aim was to create an automated system that would maintain registers of different areas. What is needed in particular is automatic transfer of the material in the building information model (i.e. the BIM material), created during construction, to a test register.
The city's employees and designers.
What were the outcomes of the experiment?
It was discovered that artificial intelligence is easy to create, but that it is not easy to standardise the process to such an extent that artificial intelligence would always be able to read it. It is not yet possible to create artificial intelligence that can read exceptions.
It was understood that a new way to make plans in a manner that is simple enough for artificial intelligence to read is required.
It was also understood that a new way of thinking is needed regarding the flow of information in the organisation. The starting point would be to find a way the information could benefit others and the entire life cycle of the building could then be considered.
A data model would perform a week's worth of work in minutes. Information can be entered into the programme only through a standard procedure, so it cannot be interpreted in more than one way.
This model frees up more draughtsmen, for example, to analyse geographic information and to do more creative work. Savings will also be made.
The test register was built using open licence codes. The instructions for Espoo's model can be found on GitHub, so anyone can see it and develop it.
Implemented by: City of Espoo
Partners: Gispo, Spatialworld Oy
Saara-Maija Pakarinen, Head of Data Management, Housing and the Environment, City of Espoo
Tel. +358 46 877 2574