Relying more on machines to do repetitive work is not new. But knowing what you want to achieve and devising a proper strategy around it can make it seem so.
“A lot of times we’ll have conversations where companies want AI (artificial intelligence) and machine learning algorithms, but they might not have started a robotic process automation (RPA) programme yet,” said Rebecca Keenan, head of process automation service for Expleo Ireland.
“With RPA, it’s a lot of repetitive mundane and existing processes – so if we can put a robot on it, why wouldn’t we? Even at the moment, I’ve seen companies where RPA can replace reactive troubleshooting with more proactive identification of the issues you’re trying to face. We’ve seen that first hand where you let a robot do a process and tell you where some of the challenges are.”
As it means different things to different people, references to AI tend to include other technologies like machine learning tools that have their own set of functions and rules. Yet, by breaking it down, you can start to see how tools like RPA can benefit sections of a business.
As well as spotting where the pain points might come from, the other benefit to having an RPA, said Keenan, is that it creates data which you can then feed into machine learning algorithms.
Much of the focus around data involves the type we’re gathering and how we’re using it. But as Jeff McCann, Director of IoT & 5G Strategy, Customer Solution Centers, Dell Technologies said there’s one crucial part to all of this.
“The data doesn’t really matter, it’s the outcome that’s the critical point,” he said. “How does it help me in my life? That’s what all of these technologies come down to.”
With tools like RPA, it’s about recognising which skills complement them best. Questions like how you manage workflows and how you help people understand the data they have, how it benefits them and how they accomplish tasks, all need to be asked. Dell Technologies itself had a major programme in place for months where its team members were trained up on RPA tools and techniques.
The crucial part of this is how the human continues to be at the centre of the decision-making process. While there are some processes that can be completely automated, others will require a human touch as there are too many variables in place.
Think about how something like Gmail would suggest an auto-completed sentence to you. It’s there if you need it, but if it doesn’t fit the context, you can continue typing what you want to say.
Keenan refers to the process of human-in-the-loop, where a machine requires human interaction, as how this works.
“That can still be that one step before AI or machine learning. It’s getting to the point where we’re building robots and they’re doing their thing. When they hit the point where they need a human decision, we can then send it back to the human,” she said.
“I get asked where I see the workplace being in X amount of years a lot, and I always answer that I still see people at the heart of it. Our offices and teams will still be made up of people; however, we will be enabled through this technology to make more informed decisions.
“We can be in a meeting and we can have real-time data informing what we’re doing. That’s where that increased innovation comes from technology. I really do see people, process and technology all coming together to be that new workplace overcoming that fear of work disruption, job loss, or maybe performance anxiety.”
Referring back to outcomes, a clear aim can create some simple yet beneficial applications for AI. While the big outcomes are the ones that capture attention, the smaller, practical ones tend to have a greater impact.
Jeff McCann, Director of IoT & 5G Strategy, Customer Solution Centers, Dell Technologies.
“A lot of people are talking about AI at a much larger scale, but it’s changing right across the board and we see it creeping into other places,” said McCann.
“To give a simple example, the latest Dell laptops now have a machine-learning engine built into them that monitors how you use your system with power and battery life for productivity purposes. Simple things like that are building on providing a better product or service for customers, understanding how the customer uses the product or service: how often do I charge my laptop? And when do I charge it?”
When you take this a step or two further, such things can be used in areas like IoT, smart cities, health and other sectors. McCann said that when a customer is looking at IoT solutions, a data set is valuable when you know what your investment and goals are. From that, you can build upon your previous work and improve future iterations of products and services.
“Say you’re a pump manufacturer. If you’re monitoring a pump, how can you improve the next generation of pumps by knowing what is happening to this generation?” McCann.
“If you take that data set that you spent time and effort making decisions on today, you can work that data back into future improvements and quality of your capabilities.”
It’s when you start thinking of AI or related technologies in terms of outcomes that processes begin to make sense. As McCann said, simple algorithms with lots of data are always going to be more accurate than a complex algorithm with less data.
Yet the crucial point to note, said Keenan, if you’re incorporating AI, RPA or any other tool into your business processes, is the importance of keeping your company in the loop.
Many implementations fail because the company kept the wider organisation out of the journey. You can scale your automation programme only as quickly as your people allow you to.
“If you’re leaving them out, that’s taking a big chunk of your organisation who aren’t giving new ideas,” Keenan said. “You need their help to learn about your processes – so, again, getting started, picking the right processes and not forgetting your people are part of the journey as well.”