AI to support the introduction of safety observations

Industry

Services

Technology

Most companies are incorporating technology into all their processes, as digital tools help improve how those processes run.

One of the aspects that must be taken care of in all companies is safety of a company since any deviation of the standards can result on serious damages for the workers.

That’s why Aquiles Solutions works together with its customers to provide them with the digital tools required to enhance their safety processes.

Approach to the problem

Safety observation is a very fruitful activity that helps find unsafe acts and unsafe conditions at the work site. Also observing and reporting of best practices followed by employees makes a difference to building safety culture across organization.

That’s why the company of this success story contacted Aquiles Solutions and requested a tool that would allow Safety Observations to be entered on the mobile in simple steps, where the user only had to enter with the keyboard, terms that were not registered in the system.

 

The Solution

It was proposed to design a predictive algorithm that would facilitate the introduction of information, asking the right questions and proposing the most likely answers, considering the data available at each time in the system.

The system would have a list of structures, with which you could know in advance the information needed in each case to complement a Safety Observation, which would be the guide when asking the questions.

 

How the tool works

For the introduction of a Safety Observation on the mobile phone, several options were selected from a list.

For example, in the case where an oil stain was observed in a specific location, the user viewed a list of possible locations (based on previous Safety Observations), and he could choose one of them.

Possible locations appeared on the list ordered by how many times they appeared in other Safety Observations.

When the process was complete, the algorithm returned the values to be entered into the tool in the different cases, which were:

  • Action type: recognition or correction.
  • Location: section and area.
  • Company/subcontractor.
  • Description of the activity and observation made (in text format).
  • Feedback (description in text format).

DOWNLOAD THE PDF

More cases

Process Partner: Anomaly Detection in manufacturing
Smart Roundabout based on Deep Reinforcement Learning
Purchase Order Reader: Automated Data Extraction