When analysing an incident, injury, near miss or observation, some findings simply can’t be determined using the human eye alone. AI has the ability to analyse insights from tens of thousands of HSE records to recognise industry-specific terminology and extract conclusions. Findings are then displayed, suggesting areas for improvement, and enabling professionals to be more proactive than reactive.
COMET Signals uses Natural Language Processing (NLP) to perform analysis on the free text description of incidents or events and determines insights such as key recurring topics or phrases, commonly occurring hazards, severity, and typical root cause categorisation. It highlights not just a mythical primary cause, but multiple causes spanning the entire organisational spectrum.
These can be presented and correlated, displaying each category in order, and revealing trends. When combining these insights with some standard classification of the events, such as business area, or location, we can start to build some powerful views of the data and address underlying issues.
AI allows your data to be tracked to show the frequency and timing of events per type, generating a comparison between periods, such as year on year. Once the data is mature, the AI can then be trained to become predictive by recognising previous insights, any actions taken and outcomes.
This all sounds great. However, the insights that can be derived can only ever be as good as the underlying data provided. Once a QHSE (Quality, Health, Safety, and Environment) professional has seen what’s possible, they will be determined that their company’s record keeping is as good as it can be, without being onerous, to ensure that real value can be derived from that data. The alternative is that this data is potentially being wasted, as no real information can be gleaned from it, so you question what is being captured and why in the first place.
This is taking the application of AI in QHSE to a different place, learning from previous events and experience, enabling teams to be proactive, and benefitting employees by mitigating serious incident risk before the bad event actually happens.