For journalists, it can be tricky to track down the right people to interview, particularly when dealing with vast, multinational corporations. And there was one company that had eluded me, until recently.
That company was IBM, and the right person, it turns out, was Sonia Van Ballaert, global client director at IBM Global Markets. Sonia is an expert in digital business transformation and leads the development of digital solutions for sustainability in heavy industries.
In the end, all it took was for a mutual acquaintance to connect us.
“It’s hard to imagine a future for the mining industry without data,” she told me, when we spoke in late September. “I think the value of data and, more broadly speaking, digital transformation for mining is huge. The industry is just starting to wake up to that potential.”
Getting down to business
Often with new digital technologies or concepts – blockchain is the prime example – there is an initial flurry of interest as to the potential they might hold. It’s usually a few years down the line before that excitement dies down and people start to understand the most effective use cases and begin to get results.
I asked Van Ballaert if that’s the case with digital transformation too.
We’ve had the technology to collect data for some time, and we know that it carries huge potential for businesses, but it’s only recently that we’ve started to understand how to make that change happen and hold the necessary skill sets. Correct?
“Correct,” she replied. “When we think about the skill sets that have emerged over the past few years, there has been a lot of emphasis on data science, as though it was the be all and end all of data.
“But if you don’t understand that data in a business context and can then cleanse and integrate the data sets properly…if you don’t have that underlying skill that we call data engineering, then the data scientists can’t do their job properly. It requires both data engineers and data scientists to work together.
“The projects around data that do succeed are, from the get-go, conceived as cross functional projects with proper business insight at the table, proper data engineering skills, and then very clever data science.
“When you combine those three things, then you will have something that leads to actionable insights you can actually work with.”
Another important element to scaling data solutions is the end user experience of that insight.
“People are not going to use your data product in the office, they’ll most likely be in the field or in an operations centre, so you need to make the product consumable, for instance, on a mobile device,” Van Ballaert explained.
“People need to trust what you’re telling them. The information needs to be presented in an understandable and reliable way on their devices and at the exact moment they need it.
“So, when you think about the value of data. It’s not just about the data, it’s about creating this entire experience of the data in the field, in the moment that is most useful for people so they can take action based on it.”
From security to skills
Of course, digital transformation can bring challenges as well as opportunities.
I asked Van Ballaert what are the main concerns that she’s seeing and how we can address them.
“Cyber security is a big one,” she said. “When equipment or processes are instrumented with sensors, the data is collected and then transferred from the operational technologies – machines, for example – outside of the operational environment into the enterprise information network, so that you can do something with the data.
“It’s known as IT/OT convergence, and this creates new hazards; opportunities for hackers to break in and even possibly start manipulating the machinery.
“The good thing is that we can out smart these threats with a holistic and integrated approach that monitors and mitigates the risk in a combination of IT/OT environments. These environments used to be very separate, whereas now you need a control room that can overlook the entire space and monitor cyber threats across the company.”
Such a security operations centre enables mining operations to maintain surveillance over all digital assets.
And what about the skills shortage that is looming? I asked. That’s another massive challenge.
“Many engineers will already have an affinity with data and math, even if they weren’t specifically trained in data engineering,” Van Ballaert said. “What’s new is the size and importance of data in their daily workflow. Handling data and alerts that come in and understanding their value is now part of the daily job.
“I think the bigger issue is in attracting new, data-skilled talent to the industry. That’s more to do with purpose and why people want to have a career in mining in the first place. What do we have to offer new generations?”
The mining industry is not alone in facing this particular hurdle. The chemicals and petroleum sectors are in a similar situation. All will need to rethink their value proposition and purpose going forward in order to connect people (especially new recruits) to their core purpose.
Edge or cloud?
Something I also wanted to ask Van Ballaert, a remnant from my deep dive into computing earlier this year, was how cloud and edge computing techniques can be used to get the most from data.
“Data underpins both of those concepts,” Van Ballaert explained. “With cloud computing, you move the data off site, into ‘the cloud’ to process it. Whereas with edge computing, the data is computed immediately within the device, very close to where it is created.
“With the emergence of 5G networks and better coverage, mines can now compute their data in situ; they don’t need to ship it elsewhere and have the results sent back.
“That means we need to start thinking about how we make the analytical applications so that they run where they need to be run. And it will also lower the cost of computing, because you don’t need to ship everything to the cloud.”
Edge computing capabilities are particularly important for autonomous mining technologies that need to process large volumes of data as quickly as possible to make safety critical decisions.
That’s why so many mining companies, particularly those that rely heavily on autonomous load and haul technologies for their business case, are currently implementing 5G networks or buying up bandwidth in preparation for installation.
And speaking of decision making… I had to ask about AI.
“AI is an often-used term, but it’s really automated learning that happens,” said Van Ballaert. “So, the machine is learning, and it can do that, maybe not as cleverly as people, but certainly much faster and more systematically. That means machine learning techniques are very good at understanding patterns in data at a rate a human would not be able to.
“Machine learning was originally applied to structured information – rows of numbers in databases – but now it can also be applied to unstructured information too, like language. For instance, to instructions written about a particular piece of equipment.”
Where I was going with my question was, of course, IBM’s Watson.
The name evokes a person rather than a technological tool and, for that reason (probably because I’m typically British and don’t like to offend anyone) I’ve never really asked or understood what Watson actually is.
Upon Googling ‘what is IBM’s Watson?’ I discovered that the company has in fact provided a simplified definition on its website (clearly, I’m not alone in my ignorance).
“Watson is IBM’s suite of enterprise-ready AI services, applications, and tooling” the site offered.
Can you tell me more? I asked Van Ballaert.
“The Watson technology is actually embedded in a lot of products that IBM sells,” she smiled. “It makes our Maximo asset management or inventory optimisation for instance really clever, giving it predictive capabilities.
“It’s also embedded in a lot of software produced by other companies who did not want to build their own AI. But Watson can also be consumed on its own, and it is brilliant at working with natural language, for instance to analyse maintenance logs or geological reports.”
But more on that later…
Data scaling & AI factories
“I’m going back to the question you asked earlier about how mining is now waking up to the potential value of data in digital transformation,” Van Ballaert changed tack. “To achieve that digital transformation, we need to scale up what used to be experiments around using data.
“There are two routes you can go down. Say, for instance, you built a pilot internally using Watson technology to understand, predict and optimise scheduling activities, then you wanted to scale that.
“You would need to do that with a long-term view because, over time, that tool needs to be maintained and there will be drift on the AI model. Maybe it learns things that are not quite right and you need to correct it or, the initial data can be augmented by other data sets to make the application even richer.
“So, when you think about the AI value chain or lifecycle, you also need to think about keeping your models and your application green and to improve and enhance them.
“It’s hard for companies to get their arms around the full software lifecycle. That’s why we introduced the concept of AI factories: it combines data, people, process, product and platform into one unit that moves beyond science experiments and delivers AI that drives business value. It combines DataOps, ModelOps, and MLOps so you can scale and maintain your AI over time.”
The other route is to buy digital tools or Software as a Service (SaaS) products and integrate them to scale up…
“Have you read about the Oren marketplace? Van Ballaert asked.
“Oil and gas companies have gone through their own digital transformation and the mining industry is actually, from a value chain point of view, quite similar. So, using that experience, Shell and IBM have partnered to create a digital marketplace with a vetted catalogue of solutions for the mining industry.”
The intent is to enable miners and providers to co-create and integrate solutions on the platform to solve mining problems.
There’s a service portfolio that comes with the platform too so Oren benefits both the sellers and the buyers, and IBM operates an open architecture policy which mitigates data ownership issues that can arise with traditional SaaS or OEM digital applications.
Data & sustainability
“The other topic I would like to mention is that of sustainability and licence to operate,” Van Ballaert said. “I believe that the future of mining, its licence to operate at all, will be determined by the trust that we can put in mining companies and their ESG [environmental social governance] track record.
“Those track records need to be proven with data, and that holds true for the carbon footprint of mining companies too.
“I think the whole ESG topic will drive the need not just for data, but for data you can trust, that you can audit, that will be reliable. The future of mining is not just autonomous, it also needs to be sustainable and for those two purposes we need data.”
To aid this, IBM has established the Responsible Sourcing Blockchain Network (RSBN) in partnership with RCS Global. RSBN ensures minerals are responsibly sourced and traced, for example from conflict regions, to battery makers and car manufacturers who use these materials.
The initiative, which has been recruiting members for a while, has now reached a critical mass which will allow it to move forward. RSC Global acts as the auditor, and the first use case is tracking cobalt for use in electronics and battery production before extending to other minerals, especially those in demand for electrification.
Next up, containerisation
I asked Van Ballaert if there are any other concepts on the horizon that hold promise for mine data management?
“There is one more and that’s containerisation,” she said. “It’s really a way of delivering applications.
“At the moment, many mines have hybrid data systems, some of which are hosted on site, some in the cloud and, in the future, they will be on different technologies as well. Containerisation encapsulates those applications and allows them to run on any infrastructure, anywhere.
“If you want to deliver an application on the edge or on different types of mobile devices, you don’t want to have to redevelop the application each time, right? You want to develop it once, encapsulate it and run it anywhere.”
In July 2019, IBM acquired software developer Red Hat, famed for its open source technologies. The company’s OpenShift technology is specifically designed to handle hybrid cloud and multi-cloud software deployments.
“If you want to deliver an application on the edge or on different types of mobile devices, you don’t want to have to redevelop the application each time, right?” Van Ballaert explained. “You want to develop it once, encapsulate it and run it anywhere.
“It’s just much more efficient and allows businesses to be more agile. Also, you can chunk up the applications into smaller parts, microservices, and then deliver the functions, much faster.”
By this point, I was way out of my depth, which I confessed to Van Ballaert. Adding that, while I don’t fully understand concepts like as containerisation, I still find them fascinating.
Pay it forward
Quantum computing is another topic that holds much intrigue for me.
In the past I have tried, and failed, to contact the experts at IBM to quiz them on quantum for industrial applications and materials science.
“I think that would be a great piece. I could put you in touch with the people who work on that,” Van Ballaert kindly offered.
And with that, we had come full circle.
The moral of the story?
Never underestimate the power of networking.