The rise of the machines: not so doom and gloom

As technology perforates every aspect of our professional and personal lives, it is clear that a robotic revolution is upon society. While robots have been used in industries such as manufacturing and automotive for years, today’s systems are ever-improving and, in some cases, are now exceeding human limitations.

For instance, Google’s artificial intelligence (AI), DeepMind AlphaGo, is beating the world’s number one player in tournaments of Go (an incredibly complex Chinese grid game). While this represents the higher end of robots’ current capabilities, there is a growing, but unfounded, fear that with the rate of development, it’s only a matter of time before all jobs are completed by machines.

Losing jobs to technology is not simply a modern day worry. For example, the early days of the industrial revolution saw everyone believing that their jobs were at risk. In the 1930s, economist John Maynard Keynes coined the term ‘technological unemployment’, believing that the rise of technology would lead to a permanent decline in the number of jobs. In both cases, the impact was not quite as dramatic. Technology helped employees to do their jobs better, it didn’t necessarily replace them. It even created new positions.

It’s a similar story today. An accountant, for instance, is likely to use tools that automate certain aspects of their role. Data collection and report creation are both arduous time-consuming activities to complete manually, and the ever-growing deluge of information only increases the chances that something vital will be omitted. Automating the processes provides accountants with the data they need to analyse, enabling them to increase accuracy and productivity.

Many will argue that automated production lines are a testament to the fact that machines can do some jobs better, and businesses do indeed invest heavily in order to have work floors almost free of humans.

>See also: VR and machine predictions for 2017

But the belief that robots will take all jobs is wrong. Semi-skilled positions will bear the biggest brunt, while low- and high-skilled jobs (such as caretakers and data scientists respectively) will be impacted less, due to cost or the complexity of roles. In fact, the rise of machines will also help stimulate employment figures by creating new roles.

Despite having all the ‘bells and whistles’, robots can’t design and service themselves. Humans can try and create another robot to do the job, but then what happens if that one breaks down too?

The truth is, the cost of developing a machine to do such a role will always outweigh paying a few humans a salary; meaning the number of positions will increase in proportion to the number of robots. Furthermore, as machines become even more intelligent there are simply more things that can go wrong, exacerbating the need for skilled workers.

Some humans will be working for or following orders from robots. In warehouses and other logistics operations, for example, machines already tell humans what to do, including which items to select off shelves and where to process them. Robots can more quickly analyse orders and delegate the relevant responsibilities to ensure that they are fulfilled in almost real-time.

While it may seem incredibly ‘overlord’, it’s worth noting that people all already work for machines in some capacity. Every time someone uses Facebook or Google, they are providing AI systems with data and they pay people back in services. People also take their driving directions from apps on their smartphones. So, just as an employee generates value for an employer, humans are all worth something to those machines.

There will also be humans who own the machines, a role that is not just going to be filled by the billionaires developing machines today. Today’s independent lorry-driver who can operate just one lorry today will, in just a few years, be able to invest in and operate several autonomous lorries.

Ultimately, while reports will continue to conclude that robots are coming for people’s jobs, it’s often forgotten just how intensively competitive humans are. Elon Musk argues that humans will have to become cyborgs to beat machines, but people compete with other people, they don’t compete with machines. Humans will always find a way to win, even if they adopt technology from machines in order to compete with each other. Even if some jobs are replaced by robots, those affected will simply re-skill themselves, and perhaps upgrade themselves, to find another profession.

Source: Information Age-The rise of the machines: not so doom and gloom


Automation doesn’t have to be a dirty word…

Without a doubt, the impact of automation on the IT Services industry is a topic of much debate and contention. The challenge is that speculation drives much of the discussion, rather than quality data and analysis.


While the subject of automation has been discussed a few times on the blog, I feel compelled to add my experiences and those of the IT professionals I met on my travels to the discussion.

Not long before joining HfS, I spent several months presenting research on automation at events and conferences across the UK. While the research covered a broad range of topics, automation in IT services was by far the most popular. After a few presentations discussing the increased adoption of automation and the growing capability of the tooling, it became apparent where the popularity of the topic originated – fear. After each session, a small gathering of IT professionals would question me on job security, headcount decreases and how automation augered a bleak future for the industry.

It’s not difficult to see why the audience felt this way. The mainstream media and even some analyst firms have been stoking the climate of fear with considerable vigor.

So I went back to the drawing board and changed my presentation. I took a fresh look at the data to examine what was happening in the industry – did we genuinely need to worry? Beginning with an impactful quote most media outlets were running with – something along the lines of “be terrified, the robots are coming” – I started to dismantle these theories with my research data on employment trends, headcount increases, and industry perception.

While many argued that automation would lead to job cuts, my data showed the opposite. Organizations recognized the importance of technology to their businesses and were investing in the services needed to support it. The data revealed that in organizations with higher levels of automation, workers were not disappearing, they were moving to higher value areas of the support structure – taking on strategic projects or developing services.

At the end of the presentation, I concluded that the reality of automation’s impact on modern IT services was far from the bleak picture painted by other analysts and consultants.

Nevertheless, a few minutes after the session ended the same horror stories started to emerge: IT leaders facing a backlash from staff as automation projects ramp up and professionals working themselves into a frenzy over their job security if projects continued. It was frightening stuff.

Crucially, my research revealed that the cause of this panic doesn’t come directly from the automation itself – there were almost no real-life examples of automation leading to sweeping changes in any of the organizations I was working with. Without a doubt, much of the fear was generated by analysts and media outlets whipping up this distorted perception, but surely there must have been another force at work.

After a bit of digging around the real cause of the hysteria became clear. In organizations with little or no perception issues, it was clear that the leadership team had taken the time to communicate with their teams. Conversely, those with stressed and worried staff had not.

When I questioned an executive who sought advice on soothing fears in his team if he had clearly explained his vision, and what the outcome of the project would be, he replied that it was obvious what he was trying to achieve. If that were true, the perception crisis in his organization would not be there.

Successful automation projects have an engaged team working behind them. The most effective I have seen understand what will be automated and why. They know what impact it will have and, for the most part, agree it was an area of manual work they found repetitive, boring and unfulfilling anyway. They eagerly anticipated a time when they could dedicate their efforts to more meaningful and valuable work.

Under different circumstances, this committed group would be dealing with the same fear and stress as their peers in organizations with less effective communication.

In the noisy information age we now live in, it’s easy to get caught up in the hype. Business leaders have an obligation to provide clear, effective communication that outlines the vision and journey of automation projects. Without the context and understanding they provide, an engaged team can quickly turn into a stressed one. And a stressed team will undoubtedly hold your project back. It’s not hard to understand why an individual afraid of becoming obsolete may not be working towards your goals with total enthusiasm.

Source: HFS-Automation doesn’t have to be a dirty word…

RPA and AI – the same but different

For a conference run by the Institute of Robotic Process Automation (IRPA), there sure was a lot of talk about Artificial Intelligence (AI). Unfortunately, most of that talk only seemed to confuse people about this latest, and most-hyped of, technologies. There were frameworks presented which showed RPA and AI as a ‘continuum’, there were models that seemed to suggest that there was a natural ‘journey’ from RPA to AI, whilst others talked about AI being a ‘must have’ if RPA was to realise its full value. Some presenters talked about a ‘choice’ between RPA or AI. None of which really helped educate the conference attendees on the benefits of either technology. Let’s unravel each of these points so that everyone can be clear on the relationship between RPA and AI.

The RPA/AI Continuum – whilst it can be argued that RPA is the relatively simpler of the two types of technologies, they are very different beasts indeed. The key difference is that the robots of RPA are ‘dumb’ whilst the AI is ‘self—learning’. The robots will do exactly what you tell them to do, and they will do it exactly the same way again and again and again. Which is perfect when you have rules-based processes where compliance and accuracy are critical. However, where there is any ambiguity, usually when the inputs into a process are unstructured (such as customer emails) or where there are very large amounts of data, then AI is the appropriate technology to use because it can manage that variability and, most importantly, get better at it over time through its own experiences. So, if you do want think of a technology continuum, make sure you put a large gap between RPA and AI.

The RPA to AI Journey – there are a number of case studies where companies have implemented RPA and then implemented AI, but only because RPA is a more mature technology than AI. There are far more examples of companies implementing RPA and not implementing AI at all because they actually don’t need the AI. RPA does a fantastic job of delivering labour arbitrage, accuracy and compliance without AI coming anywhere near it. And, of course, some companies implement AI without RPA. It’s not a journey, just a set of choices based on specific demands.

The RPA Dependency on AI – another view that was put forward was that RPA is only valuable when it has AI in support. This is clearly a self-fulfilling view put forward by the vendors that are able to offer both technologies, but it is simply not correct. As mentioned above, many (in fact, most) companies implement RPA without any consideration or need for AI. If you want compliant, repeatable processes, and can feed the robots with structured data, then why complicate and confuse matters by introducing AI?

The RPA/AI Choice – There was yet another the view put forward (which actually conflicts with much of the above) that companies need to make a choice between RPA and AI – in other words which is the best one for them to implement that will deliver their objectives? As should be clear by now, the two technologies actually complement each other very well, for example by using AI to structure unstructured data at the beginning of the process, by using the robots to process the transactions, and then potentially using AI for decision making and/or data analytics at the end.

So, why all this confusion and mis-information? Part of it is obviously self-interest from vendors and providers to create frameworks and models that align with their own capabilities and marketing messages. And, although RPA is now pretty well defined (with that badge of maturity: its own acronym) some of the confusion surely arises from the multiple terms used to describe artificial intelligence; AI, cognitive computing, machine learning, NLP, etc. For now, it is much the best approach to think of AI in terms of how it can help your business, without worrying about what to call it. As the technology develops though a more robust approach is required, which is why at Symphony Ventures we are working on an ‘AI taxonomy’ that will clarify the different types of AI, and therefore help to explain the practical opportunities and uses for AI in our clients. We look forward to sharing this with you and de-bunking much of the confusion around RPA and AI that we have seen over the past few months.

Source: and AI – the same but different 

The Key to AI is Structured Data, BNY Says

The promise of artificial intelligence around streamlining and enhancing banking processes has enticed most, if not all, financial institutions at this point.

But before jumping to AI, FIs need to get their data straight, according to Jon Theuerkauf, managing director and group head of performance excellence at BNY Mellon.

“Forget AI, I don’t even know what it means,” he said at BluePrism World event today. “Why are we jumping on it, if we haven’t done the basics?”

The “basics,” according to Theuerkauf, is having a structured data ecosystem in place.

“We are now in a transitional phase, and are still three to five years away from integrating operating automated environment,” he said. “For example, it takes a long time to train Watson. Why? Because the data does not land itself easily to allow Watson to learn. So, there needs to be an order around that data, and we are now starting to put things together and taking the chaos out of it.”

Currently, BNY is “consciously incompetent” in terms of adopting AI and automation, and “we are really proud of that,” Theuerkauf added. “We now know what we don’t know, we know that we need ways to structure data, and are still working on how.”

BNY began experimenting with bots and automation last year. In less than 10 months, the bank went from 0 to 200 bots, according to Theuerkauf, and has thus far automated more than 100 processes, with 50 more underway.

“We have $3 trillion in payments a day that run through our shop, would you let a robot handle that process? This was actually the second process we had automated,” he said.

The efforts around structure and RPA — robotic process automation — will, eventually, lead the bank to AI and machine learning integrations. “We are figuring stuff out, putting teams together, and hopefully this will lead us to maturing into a machine learning and AI-enabled world,” Theuerkauf said. “RPA is the ‘doing’ part of automation, AI is the ‘thinking’ part.”

Source: Key to AI is Structured Data, BNY Says