Kaios.ai uses artificial intelligence (AI) to enrich data and automate processes. Today we will be discussing the case of Asfalt-onderhoud. Asfalt-onderhoud is a company that does small-scale maintenance on asphalt. They repair cracks and holes in roads, bicycle paths, footpaths and playgrounds. The total repair time determines their billable costs.
Before they started using the AI-solution Kaios.ai created for them, they manually checked video recordings of their completed maintenance. This was a very time-consuming process, and for that: a perfect case for Kaios.ai! Because Asfalt-onderhoud already did their own data collection (making video recordings of the completed maintenance), Kaios.ai could quickly start to develop an AI-solution that works with the collected data.
For the development of the solution, Kaios.ai and Asfalt-onderhoud work closely together. Mickel Koeslag, planner at Asfalt-onderhoud, explains that Asfalt-onderhoud is a very innovative company whom already acknowledged the problem of the inefficient process of the annotation of the completed maintenance. That is why they switched their measuring process from a measuring wheel to making video recordings of the completed maintenance. So, before Kaios.ai was consulted, Asfalt-onderhoud was already busy optimizing their annotation process.
Asfalt-onderhoud is very much into trying new things, but they do want every improvement to be easy and efficient. He emphasizes their business philosophy as: (It has to be) easy to understand, quick to work with and effortless to operate. In the particular case of improving the annotation process of the completed maintenance, the biggest challenge was to make it so called ‘idiot proof’. So, for that part: how did Kaios.ai and Asfalt-onderhoud came to the AI-solution they are using as we speak?
Mickel clarifies that Kaios.ai starts building their AI-solution after Asfalt-onderhoud had explained their annotation issues. The process of testing the solution and giving feedback on the usage is the most important part in to creating the best AI-solution. This feedback mostly comes from the employees of Asfalt-onderhoud using the solution during their work. Mickel makes sure he gets feedback from his team, bundles it all together and communicates it with Kaios.ai.
The outcome of this feedback could sometimes be very practical. For example: before the implementation of the AI-solution, a video recording of the completed maintenance was made after every damage repair. Now, only one video recording is made, after all maintenance is completed. This ensures that the AI-solution only has to process one smooth recording, instead of all separate recordings but this change also speeded up the work process of Asfalt-onderhoud.
By giving feedback and input back and forth, the AI-solution starts to evolve into the desired product. The two companies will stay in contact. There are several new ideas to discuss and it’s important to have the solution to move along with potential wishes or changes at the side of Asfalt-onderhoud. Mickel explains that building an AI-solution is a continuous process during which Kaios.ai and Asfalt-onderhoud are in contact at least every two weeks. By fixing one problem, another one arises, so it is important to always continue further development.
Curious about this solution? Check out our video.