Though the world is rapidly evolving with advances in AI and other technologies, the H&S sector is playing catch-up. It is relatively under-digitised, less connected, and has sometimes been slow to embrace meaningful digital transformation. It’s a highly competitive industry, on the cusp of a revolution driven by wide scale adoption of AI.
Innovative stakeholders are finding new ways to use AI throughout their workflows right through to the job site. The most meaningful examples are the ones that improve human outcomes while making proper considerations for ethics and responsibility.
Perhaps the most important of all problems, the one that weighs most heavily on the hearts of those who have seen it go wrong is this:
How can I make this safer for the people I care about?
It’s a question asked by heavy industry and Health and Safety Organisations alike. They’ve developed practices that have significantly reduced the number of fatal and non-fatal incidents so much over the last few years, that the number of fatalities reported is beginning to plateau. This is great news and suggests that the most obvious patterns of hazardous conduct have been rooted out.
But what about the less obvious hazards?
Hazards are becoming increasingly more difficult to find and therefore more challenging to target and prevent. AI and machine learning provides an inroad here; leveraging the power of data at scale to assist humans in driving down accident rates even further.
The challenges to achieve this are varied. From disconnected, unstructured and incomplete data to a huge range of working landscapes with job idiosyncrasies and nuanced context in abundance, many aspects of the health and safety landscape conspire to ensure that no one-size-fits-all approach to unlocking the value of the data will succeed.
These data formats are often simply not parsable by an automated system without an understanding of data context, interconnectivities and latent structure. Equally, the nuances, constraints and peculiarities of individual settings such as job sites, workforces and other environmental conditions are often not represented in the data, meaning that any automated model will gloss over these aspects, at best leaving a large amount of predictive power on the table, and at worst providing outputs which are not applicable to the target setting.
Mind Foundry’s technology explicitly shuns automated black box systems, instead focusing on human-AI collaboration driven workflows to get the best out of the data and leverage the domain expertise of health and safety professionals. The Mind Foundry Platform provides an iterative workflow allowing H&S professionals to rapidly get AI-driven results, in conjunction with their domain expertise. Results are improved collaboratively by incorporating new data sources, confirming and refining AI derived features and links, and optimising workflows. This collaboration with intelligent systems delivers more targeted, predictive and applicable outcomes for specific jobs, sites and workforces.
Mind Foundry’s approach has helped H&S functions across a number of verticals to make significant protective interventions in their work practices.
For example, one of our clients found that certain types of accidents, like lacerations and eye injuries, have a clear spike of occurrence at specific times of the year. This inspired them to reschedule their employee training programme to target the prevention of these types of injuries to a time of year when it would be most relevant.
Independently, when analysing the outcomes of inspector visits to construction sites, they realised that joint-venture projects had very different practices than projects where their construction company was the main contractor. This allowed them to reconfigure and better target interventions such as inspections by the site inspector and efficiently reduce hazards. The client reported a net increase in profits given a reduction in time off due to accidents. This focus on employee wellbeing rather than blindly driving down costs delivered a better outcome for both metrics.
The Mind Foundry Platform has also been used to quantify damage and safety risk associated with work tickets in a number of settings. Again, the collaborative workflow enabled domain experts to teach the system to extract information from unstructured data, in turn identifying previously undiscovered correlations, and better predicting the riskiest jobs.
Crucially, Mind Foundry systems do not automate away the human. We understand your expertise is not encapsulated in data. We enable you to add human expertise to deliver a safer work environment and game-changing results.