The links between the Internet of Things (IoT) and artificial intelligence (AI) are irresistible. As a Wired article put it, IoT ‘will produce a treasure trove of big data’ and that ‘the only way to keep up with [it] and gain the hidden insight it holds is with machine learning.’
For Johan Krebbers, IT CTO at Shell, it’s non-negotiable. “IoT on its own is useless,” he explains. “It only becomes useful when we pick up the data, and apply machine learning to it, digitise it, and help the decision maker.”
Founded in 1907 Shell, like many of the more innovative companies who have been in business for more than a century – think General Electric as another example – have been doing IoT for many years. It wasn’t called the IoT back then, of course; among Shell’s earliest innovations was providing underwater robots back in the 1970s.
Yet the rise of machine learning in recent times – in order to facilitate real-time insights, as well as provide self-learning mechanisms – has taken the next step up. “We have to make sure we’ve got the whole end to end workflow in place,” adds Krebbers. “To just talk about IoT itself is a waste of time. IoT by itself only does data collection – you need all of the things to come [together] for your decision making.”
It all sounds great in theory. However many organisations are struggling to take the plunge. Back in 2016, Maciej Kranz, vice president at Cisco, published Building the Internet of Things, a guide for companies to begin their IoT journeys, outlining the business value proposition, organisational and cultural changes, and how to avoid mistakes.
Earlier this year, Kranz published an additional workbook to give further assistance. Under the section of assessing technology readiness, organisations have to answer whether they would be able to connect and access all data to ensure it flows to all. “IoT does not require the latest in technology,” the book asserts. “You can start simply by connecting together the systems you already have. That alone will generate new value.”
But what if your data is not of good quality? Krebbers certainly does not see it as an issue. “I’m less worried about quality of data because you can never get good quality of data,” he says. “If you wait for that, you wait until the cows come home, so you can’t wait for it.
“You have to use the data you have today, start using it, make it visible, and then start improving,” adds Krebbers. “Whatever data you have today, start using it immediately. Otherwise you can spend the rest of your life improving data quality… we’ve done that for the last 500 years, and we’ve not succeeded.”
This is worth noting particularly given how complex the oil and gas industry is. As one company trying to disrupt the industry explained to me last year, oil and gas is ‘the A to Z of industrial process in one business’, taking in water, engineering, infrastructure, distribution and more. The point still stands however.
“It’s complex, but you need to get started,” says Krebbers. “You can’t use poor quality data as a reason not to do it – that’s a very poor excuse.”
The message is simple therefore. It’s no longer enough to tiptoe around these emerging technologies – including blockchain, which Krebbers says will provide a ‘huge impact’ on the enterprise – but instead get going. And the results can be stunning – Shell Nigeria has saved more than $1 million by using sensors to monitor oil fields.
Krebbers will be keynoting at IoT Tech Expo Europe in Amsterdam next week on how IoT is a ‘data driver for AI’ – and it will certainly be a session to look out for. “You need to look at the end to end story,” he adds. “You can start today, because many companies already have IoT today, you don’t need fancy stuff – and of course data is at the centre of all this.”
You can find out more about the IoT Tech Expo Europe here.