Nearly two-thirds of the global population experiences severe water scarcity for at least one month each year, while companies reporting to CDP estimate over US$531 billion in potential water-related financial impacts. In Africa, climate variability and ageing infrastructure make water risk a critical priority for organisations across sectors.
According to Alastair Bovim, CEO and co-founder of environmental intelligence platform company, Insight Terra, the ability to see how water systems are behaving in real time is becoming critical to managing those risks. “Water is the lifeblood of industrial operations, and you cannot manage what you cannot see,” he says. “But many organisations are still managing it through periodic reports rather than continuous oversight.”
The visibility gap
In many operational environments, risk management still relies heavily on periodic reports, manual checks, and fragmented datasets. These approaches provide only a snapshot in time and often fail to capture how environmental systems are changing.
”You wouldn’t want your surgeon reading last quarter’s vital signs to assess your condition on the operating table today,” says Bovim. “The same logic applies to water systems. By the time a periodic report lands, conditions have already changed.”
Continuous monitoring closes that gap – and in doing so, quickly surfaces issues that periodic reporting simply cannot catch: instruments that have gone offline, communications failures interrupting data flow, sensors that are faulty or misconfigured.
These gaps may appear minor in isolation, but left undetected, they create blind spots where larger risks can compound unnoticed.
“In our experience, the biggest risk exposure rarely begins with a complex engineering failure,” says Bovim. “More often, the trigger is a simple operational gap.”
Fix the basics first
Addressing these issues is not necessarily complex, but it does require consistent oversight.
“Once organisations move from occasional reporting to continuous monitoring, the first step is usually quite practical,” Bovim says. “You fix the basics, restore the data flow, and remove the blind spots. Only then can engineers and operators focus on the genuinely difficult problems and understand what is really going on in their assets.”
Often, continuous monitoring quickly reveals that many instruments are not actually reporting reliable data or are simply offline.
“We have arrived at facilities to find sensors that look perfectly operational – powered, mounted, undamaged – but have been transmitting nothing for weeks. The SIM card had run out of airtime. Not a technical failure, not an engineering problem – a prepaid data contract that nobody noticed had lapsed. It costs less to fix than a tank of fuel. But without continuous visibility across your monitoring network, you would never know it had stopped,” says Bovim.
Mining as a proving ground
Mining offers a sharp illustration of what is at stake. Water underpins virtually every stage of mineral processing, from ore separation to dust suppression, and its mismanagement carries consequences that extend well beyond the mine fence. In water-stressed regions, operations and surrounding communities often draw from the same sources – making water stewardship both an operational and a social licence imperative. You need to manage your water like your money: know where it is, where it is going, and whether the numbers add up.
In Chile, Insight Terra has worked with Anglo American’s El Soldado mine, where a demonstration facility is recovering and recycling water directly from the tailings system. Clean water is returned to the processing plant, reducing the need to draw additional freshwater in one of the world’s most water-scarce mining regions. Continuous monitoring helps engineers track how water moves through the system and respond quickly if conditions change.
In Southern Africa, shifting rainfall patterns are amplifying risks at tailings storage facilities and opencast operations, where water accumulation can move from manageable to critical within hours.
“The pressure is coming from both ends,” says Bovim. “Less reliable rainfall, and higher scrutiny of every litre used. Operations that lack real-time water visibility are flying blind in conditions that no longer forgive it. Unlike electricity shortages, which can sometimes be offset with alternative generation, water scarcity is far harder to solve quickly. Once supplies are constrained, recovery depends on rainfall, recharge, or expensive infrastructure.”
What has changed is the sophistication of the tools available to do this. Satellite observations, field sensors and AI-driven data platforms now make it possible to ingest, normalise and continuously analyse multiple data streams from distributed assets – flagging anomalies, predicting risk conditions, and feeding decision-makers with the right signal at the right time. The constraint is no longer access to data. It ensures that data is clean, continuous, and connected to models that can act on it. “AI is only as good as the data flowing into it,” says Bovim. “Get the data infrastructure right, and the analytical capability compounds. Get it wrong, and the models are working with noise.”
“These tools are not about replacing engineering judgement,” Bovim says. “They are about making sure that judgment is applied to accurate, timely information – not to data that is three months old or riddled with gaps from sensors that were never actually transmitting.”
A practical approach to resilience
Technology is an enabler, not a silver bullet - and this point matters more than it might appear. Speaking at the SAIMM Tailings and Mine Waste Conference, Bovim drew on two decades of high-consequence systems analysis to argue that the most persistent risk in industrial operations is not equipment failure. It is the gradual process by which early warning signals are seen, documented, and then quietly absorbed into a new definition of normal.
Under sustained production pressure, organisations learn, almost imperceptibly, to tolerate conditions they once would have escalated. By the time a critical threshold is crossed, the decision-making chain that should have responded may have already stopped functioning.
This is why continuous monitoring and AI-driven early warning are necessary, but not sufficient on their own. The signal still has to reach someone with both the authority and the willingness to act.
“We can instrument everything,” Bovim says. “We can flag anomalies in real time, connect sensors to digital twins, feed models with live data. But if the person who sees that alert has learned – consciously or not – that raising concerns has no consequence, the technology has not solved the problem. Resilience is built when the data infrastructure and the human response system work together.”
“In a climate-volatile and resource-stressed world, the organisations that will build lasting resilience will treat data reliability and human accountability as two sides of the same problem,” Bovim concludes. “Fix the basics, ensure the signals are getting through, and build a culture where acting on them is expected, not exceptional. That is what separates operations that compound their advantage over time from those simply waiting for the next incident.”