Zero harm is a universal aspiration. But you can be perfectly safe and go perfectly broke. How can we ensure the best chance of achieving zero harm while providing a sustainable and competitive return to our shareholders?
This is Part 4 of a series. Read the other parts here:
[ Listen to the audio version, read by David Hodes]
Boards must ensure that the corporations they govern have every chance of keeping their staff safe for the best life they can lead. Injuries affect employees, their families and their colleagues; fatalities are devastating. A high injury rate stains both the organisation and the reputations of individual leaders who failed to prevent such incidents from occurring. It is unethical not to care about the safety and wellbeing of the people who work with you and for you, not to mention the criminal liability that accrues up the chain of accountability when occupational health and safety systems fail.
But, like all things in life, safety is not absolute. People are injured, people are maimed, and people are killed at work. If you wanted to keep people perfectly safe in industrial environments, you’d have to stop doing work. It’s no different to thinking that if you wish for no car accidents, then take all cars off the road. But, if we don’t have anyone turning up to work because of the risk of an accident, our companies and whole economies would quickly grind to a halt.
So how do we come as close as possible to delivering zero-harm workplaces when we need to remain competitive? If we want to win over customers, we must demonstrate that we can win on price and availability without taking expedient shortcuts with our ethics. As employers, we must be fair dinkum regarding the value we place on our employees’ occupational health and safety.
Let’s start by framing how we think about safety practices at work as something we do to keep us safe for the best life we can lead—whether at work, at home, or in the community—rather than safe from dangerous equipment, hazardous environments, or risky practices. In doing so, we provide more motive power to enterprise-wide efforts to achieve zero harm. We must put the individual—where the worker meets the work—in the best possible position to be safe to live their best life. The individual worker becomes the locus around which all health and safety efforts are organised.
“The individual worker becomes the locus around which all health and safety efforts are organised.”
From the place where the worker meets their work, enormous effort has been applied to build out systems of safety that scale up through supervisors, superintendents, managers, general managers, vice presidents, and presidents. Processes are developed to incorporate safety protocols, and controls are put in place for every conceivable activity one might undertake that has a safety element. Consultants and government agencies help to build out and audit the systems. Colleagues are encouraged to look out for each other, and slogans such as ‘we all guarantee everyone goes home safe and well’ help to spread the word that safety is everyone’s business.
To make these controls come alive and gain increasing relevance, leaders try to bring safety awareness into the everyday habits of the workplace by, for example, starting all meetings with a safety moment. Posters are placed on the walls urging you to hold a handrail when climbing or descending the staircase, and disciplinary action can be taken even if you are caught outside work crossing a public road against the lights.
Senior leaders most want to achieve three things: that processes are fit for the purpose for which they have been designed; that everyone is fully present and aware to doing their work in such a way that they can come as close as possible to the aspiration of zero harm; and finally, that the cost impost of these health and safety controls are not so burdensome that it robs the organisation of their ability to sustain their competitive edge.
Let’s look at those needs and see if the advent of artificial intelligence (AI) and machine learning (ML) can help make a material difference in accomplishing superior health and safety outcomes while simultaneously lowering the cost to the business of doing so. In other words, can we square the circle of safety and productivity?
One of the issues surrounding safety controls is that they vary over time between different parts of the organisation. Many large corporations, in whose head office ultimate accountability for safety resides, have multiple divisions, often the result of mergers or acquisitions. Each has brought to their health and safety practices their own way they run controls and generate documentation, making it very difficult to make comparisons and see patterns across the enterprise.
When it comes to analysing incidents, a Pareto analysis could yield critical information about which controls are working and which need improvement. But, too often, incidents are recorded in various ways. Comparing apples with oranges means you don’t know which controls are effective in reducing what type of incident.
“Comparing apples with oranges means you don’t know which controls are effective in reducing what type of incident.”
Addressing this problem of standardising safety controls is the first place where AI and ML can make a huge difference. With the advent of natural language engines, it is possible to import the original control documents into the system and have the AI agent return results that are normalised across all instances. In other words, the AI engine looks for recurring patterns of words and phrases and ‘learns’ how to group them.
So, for example, if there are 100 documents across the organisation which address how best to be safe when working at heights, you can relatively quickly create a master document that encompasses everything unique and nothing that is duplicated when it comes to working at heights. In one case I’m associated with, the AI engine has addressed over 10,000 separate documents to create a single, normalised corporate library of unique safety controls.
The next step is to normalise all the incidents. We humans can recognise that when we read a report that says, ‘meter reader bitten by the dog’, it’s the same as saying, ‘the dog bit the meter reader’. However, trying to query these results traditionally wouldn’t work, as the data is not sufficiently structured to do so. With an AI agent, these problems can be relatively easily solved, right down to the level where you know that 2 pm on Wednesdays is not a good time for meter readers to wander through suburbia unless they take the risk mitigation step of carrying around a packet of dog biscuits.
Once the controls have been reduced to a standardised set and incidents are recorded in a standard way, it becomes possible to address the first of the safety prerequisites demanded by executives: that the processes they have are fit for the purpose for which they have been designed. One can see which controls have affected which incidents. A method such as the 5-Step FOCUS can be used to repeatedly zero in on the actions necessary to eliminate the worst offenders and move systematically to the next and the next, ad infinitum.
But how do we deal with the second prerequisite: that everyone is fully present and aware to doing their work in such a way that they can come as close as possible to the aspiration of zero harm? Once again, AI and ML can help. It is of the nature of human learning that we forget what we’ve learned. Different people have different ‘forgetting curves’ for different topics, and we all see the sharpness of our recall deteriorate over time.
“It’s human nature to forget what we’ve learned.”
One way we can deal with this forgetting is to adopt the use of spaced repetition of the learning. Let’s say you have been through a site safety induction, which included some learning modules along with an assessment. Some weeks later, you are asked to recall something specific from that induction, and you have forgotten. The AI-powered system learns what you forget, comes to know the time duration between your last test and your forgetting, and serves up targeted microlearning refresher courses for you to complete before the following assessment. What the AI agent can do that a traditional spaced-repetition learning system cannot is to tailor the intervals and the topics precisely to your needs.
But that’s not all. We can connect the worker to where they meet the work, for example, through a mobile notification of an upcoming job. Then, we can attach to the work order all the safety controls necessary to be fully kitted to complete it without injury. At a suitable time before the job starts, the worker can be assessed on the steps needed to safely complete the job, considering their personalised forgetting curve about the topics at hand.
At the next level up, the supervisor can review a ‘heat map’ of their team’s readiness to safely perform the job, either in advance or at the work front. Let’s say the supervisor had two boilermakers, a rigger, a fitter and a couple of trade assistants in their team. With individual assessments for each team member, and an automated check of their formal qualifications, the supervisor could easily see significant risks areas. So, for example, let’s say that the job involved working at heights in a confined space. The supervisor looks at the heat map of the knowledge in the team and sees that it’s red (bad) when it comes to working at heights and blue (good) when it comes to confined space. They then know what additional training and coaching they need to make sure everyone goes home safe.
The constantly updated information on every employee can be used proactively in assembling teams. Let’s say we have someone who is strong on a particular safety control required for a given piece of work. The team also has a novice who is yet to learn the correct way to stay safe on a job of that nature. The system can ensure that when the team is put together, the novice is always paired with a master to ensure the transmission of good health and safety habits.
And, of course, the assessments and spaced-repetition learning doesn’t stop at the level of the supervisor. All the way up the management accountability hierarchy, microlearning courses, spaced repetition of learning, and appropriate assessments can be developed that help superintendents, managers, and executives understand, at their level, their legal and leadership obligations.
What about the third prerequisite: that the cost impost of these health and safety controls are not so burdensome that it robs the organisation of their ability to sustain their competitive edge? Using such a system that zeroes in on the individual worker meeting their work and its task-relevant safety knowledge within the horizon of their personal and empirically derived forgetting curve must mean that the worker is far less likely to get injured.
Finally, when targeted and efficient alternatives are available, you don’t have to take people out of work to give unfocused refresher courses and validations of competency. Best of all, through such a system, we learn how to focus on those safety protocols which make the most significant difference to the most consequential of incidents, thus foregoing the vastly more costly and less effective option of the shotgun approach. Which ensures we are focusing on the constraint.
This is Part 4 of a series. Read the other parts here:
The change to using Theory of Constraints (TOC) as an underlying operating system is both profound and exhilarating. We’ve developed the Systems Thinker Course to bring the ideas into your organisation.
Few performance standards deliver the competitive advantage you gain by keeping your promise to deliver on time, doing so faster than your competitors, and suffering no defects while you’re about it.(more…)
Eli Goldratt famously said, ‘Tell me how you measure me, and I will tell you how I will behave. If you measure me in an illogical way… do not complain about illogical behaviour.’ If you measure and reward activity, then activity’s what you’ll get.(more…)
Discover better ways to do better work.
We alternate our own actionable articles with three relevant links from other authorities.We’ll only use your email address for this newsletter. No sales calls