Back to Basics is a weekly feature that highlights important but possibly overlooked information that any EHS professional should know. This week, we examine automation, the use of automation in EHS reporting, and how it can increase the productivity of the workforce.
New technologies are constantly being invented and distributed to consumers, including versions of artificial intelligence that are used in various ways to make people’s lives simpler. In the EHS industry, there has been a surge of individuals and organizations who are advocating for the automation of EHS processes, especially when it comes to reporting and data collection.
Many companies have come up with their own algorithms and data input resources to help EHS professionals with their jobs, and as this continues to happen, it’s becoming increasingly important to understand what automated EHS means and how it can impact the workforce. Jake Freivald, the Vice President of Fulcrum, spoke in a recent webinar about the benefits of automation in EHS and what the implementation of artificial intelligence would look like, specifically in the case of safety inspections.
Workflows and processes
In order to increase productivity and efficiency, EHS leaders should first consider the processes that their employees go through while doing their jobs. Freivald said there are two levels of processes, the worker’s process, and the enterprise’s process. The worker, an EHS professional in Freivald’s example, may send out a team to do an inspection on-site with checklists, then someone has to return to transcribe that data into a spreadsheet, which is then used as a part of a master spreadsheet that is turned into a report for stakeholders.
This workflow fits into the larger enterprise process, which can involve using the spreadsheets to bill clients and making sure the work being done is remaining in compliance. Freivald stated that in this framework, the data is only useful depending on the quality, which is why it is important to consider the workflow to see if there are any gaps or errors in reporting that can be potentially addressed by automation.
Automation basics
There are many different ways to introduce automation into a workflow or process, and it’s important to have realistic expectations of what automation can do. Freivald said that rather than artificial intelligence (AI) making every process magically easier, the reality is that AI will likely have a specific set of capabilities that are a solution for a single problem, and it will be able to solve that problem faster than a human.
There are multiple pathways that employ automation and technology that Freivald outlined that are specific to the EHS industry.
- Planned inspections: a manger uses a desktop to schedule the inspection, type, and the inspector, the actual inspector uses a phone to record the inspection result data after performing the inspection, and then the manager tracks inspection completion and reports the status of the inspection.
- Hazard observations: a team member uses a phone to record observation data, a superintendent uses a phone to triage the observations and add remediation, and then a remediator uses a phone to record their fixes.
- Push notifications: the inspection team admin creates a workflow that requires sending a text when the inspection status is updated, the inspector either passes or fails and has access to the inspection data on their phone, and the new automated system sends a text upon a passed or failed inspection that can provide further necessary details, like the location of a hazard.
- External data system: a geographic information system (GIS) admin plans and analyzes project maps and layers, the inspection team carries out the inspection and adds data to the cloud, and then the GIS admin requests the inspection data from the cloud to make the necessary updates.
In terms of digitizing inspection processes, Freivald recommended finding methods that make it easier for users to input data. When picking between fill-in-the-blank or a pick list with options, choose the pick list. If the worker needs to type in their observations, add in an audio feature so they can use their voice to record what they need to. Break up text and questions into appropriate chunks, and only display certain information or questions when the circumstances make it necessary. Lastly, include diagrams and any other data the user needs in the inspection checklists, and generally simplify more than what might seem necessary. The goal is to streamline processes to increase efficiency.
Increasing productivity
The first step to increasing productivity with automation is to find parts of the workflow or process where there is a significant amount of data passing between people. Freivald recommended looking for areas where poor data quality has caused significant issues, because automating those data handoffs will reduce transcription errors which makes reporting more efficient and accurate. However, it is important to find a balance when determining which problem to tackle with automation. If the problem or process is too simple, the outcome will not be significant enough for the results to matter, but avoid trying to take on too much at once.
Freivald used remediation as an example, which is a process that basically consists of identifying the problem, communicating it, fixing it, and then communicating the fix. Typically, inspectors will perform inspections, identify issues, and then transcribe the data they found in a report which gets sent to the superintendent. The superintendent will identify issues in the report and contact the remediation crew. Remediators will fix what they need to and make another report, which goes back to the superintendent, who then confirms the remediation and closes out the case. Freivald stated that this process would be easier if automation was used during the first inspection step with data collection, since it would reduce transcription errors and streamline the reporting process.
Outcomes
In terms of measuring success, Freivald said to leverage data flow to create and publish metrics, and he suggested using two different types – safety performance and inspection performance. Safety performance metrics can include remediation time and cost, fines, calendar days lost, and person-days lost. Inspection performance metrics can include inspection time and cost, the time workers wait for inspection completion, and safety problems found post-inspection. Freivald also warned against being punitive, and emphasized that punishing hazard reports teaches people not to report hazards, which is not the ideal outcome.
Finally, in order to convince upper management to invest more in automation, AI, and technology in general, EHS leaders should emphasize the end result of more productivity and the downstream effects of that productivity. Freivald said the number one thing to do is focus on specific problems that upper management sees, not just the niche issues. Frame the conversation around how automating inspections or other processes with reduce the number of billing days and the days it takes to send the bill to clients, which will increase the speed at which the company gets paid.