Humanising the Future of Work podcast has been saved
Humanising the Future of Work podcast
Bonus episode 3: Potential: AI and superteams
In this episode our speakers take a look at the second paradox of this year’s trend report – Potential, exploring security in a world of reinvention. What impact is AI and automation having on work, the workforce and workplace? Our speakers discuss how AI can be used to establish consistency, increase productivity and unlock value in teams. Have a listen!
Gloria Viedma Navarro
Host: Sam Shindler-Glass
Oscar Hamilton, Director and FSI Digital Finance Transformation Lead, Gloria Viedma Navarro, Senior Consultant
Welcome to the bonus episodes of the Humanising the Future of Work Podcast series. During these sessions, we’ll be taking a deep dive into Deloitte’s 2020 Global Human Capital Trends report, focusing on the trends we expect will have the biggest impact across work, workforce, and workplace.
Hello and welcome to our third HC Trends podcast, where we’ll be talking about the second paradox in this year’s trends report, potential, which looks at security in a world of reinvention. I’m Sam Shindler-Glass, and I’m delighted to be joined by Oscar Hamilton and Gloria Viedma Navarro from our robotics and intelligent automation practice. So, Oscar, AI has often been talked about as the future, but it’s now arriving. So, what’s next?
OH: Thanks, Sam. Firstly, I think AI itself is a broad term. AI really refers to human-like intelligence in machines, and that itself is a spectrum of capabilities. Within AI, organisations have been moving up the spectrum in the last ten years at different speeds. So, many have started on the journey with RPA, moving into cognitive technologies to understand documents, sight, voice or process.
Deep learning, neural networks, part of AI, are less taken up by organisations and primarily used for some specific use cases.
GV: Yes, actually, if I can add on to that. I think it’s quite interesting that we’re talking about the spectrum, and we’re talking about AI and what’s next. But I think it’s worth highlighting that some organisations are even struggling with more basic forms of automation, such as RPA. We run an annual survey around automation. And last year, on 2019, we actually saw that only 8% of organisations have achieved scale with RPA. When we talk about scale, we’re talking about 50 automations or more that are in production.
And so, I think this also shows that although some organisations are really going into the innovation space and trying to explore with new technologies, there is still a lot of work to be done to really overcome those barriers that organisations are encountering when trying to even just nail the basics, so to say. So, I think that’s quite an interesting point to make around that.
OH: Thanks, Gloria. Also, as the adoption of AI evolves over time, you fast forward, we’ll increasingly harness the power and insight that data can give you, and data using the cloud.
A bit further out, this will really accelerate when compute power reaches new levels with quantum computing one day, and it’s really exciting to think about the possibilities if you combine this deep, rich data set that automation, that AI can provide, create a digital copy or twin of the organisation, and then use that to run simulations to find the optimum path, in the same way, basically, that Formula One teams simulate different scenarios to plan their race strategy.
Also, what’s next? I think we’ll see more practical uses of blockchain. And then further out, really increasing human augmentation. So, if you’ve read anything around Elon Musk’s Neuralink, that’s basically an ultra-high bandwidth brain machine interface to connect humans and computers.
SS: So, we talked there about moving up the spectrum and the range of capability. Obviously, AI is becoming more and more capable. So, what does it mean for humans? And a lot of the talk for automation has led to job losses. Does it look like this is going to be the case?
OH: I think you really need to stand back and look at history. We’re in what we refer to now as the Fourth Industrial Revolution. Every previous industrial revolution involved some form of automation of fairly giant swathes of the workforce. Really importantly, what every one of those industrial revolutions has done is also created more jobs than were automated.
In this Fourth Industrial Revolution, automation is fundamentally changing, actually, the way we work, and it’s creating a need for new skills and new capabilities. So, clearly, with automation comes the automation of some tasks and some roles. RPA basically automates human robot-like tasks. So, those more manual, repetitive, rules-based tasks that are typically associated with the least skills and the lowest levels of job satisfaction in the workplace.
What is replaced by that is the emergence of new roles and new skills. So, the robot controllers. So, you run and control and orchestrate robots. Also, data scientists, and engineers, and AI capability.
GV: Yes. And if I can add to that as well. I think what’s really important is that workforce transformation efforts are crucial. They need to go hand-in-hand with wider digitisation efforts. So, when an organisation is driving their intelligent automation programme or is exploring new technologies, it’s really, really important that that’s taken into consideration with the wider workforce reskilling and upskilling that will result with the digitisation effort.
So, I think the workforce needs to be prepared for working in a completely different way. When we implement automation, we’re essentially changing the way work gets done. The way teams are structured, but also the capabilities and skillsets that will be required in the organisation to drive that future is very different.
So, I like to think about automation as it’s augmenting the human workforce. It’s helping to move up that value chain of work and allow humans to do more value-add work. Leave that repetitive, less value-add tasks to the bots, and allow humans to be what they’re best at. Be creative. Be social. Be empathetic.
But also, linking back to the point that Oscar made. It’s true that when you as an organisation start scaling those automation initiatives or digitisation projects, it will create new job families. So, new roles emerge across the organisation that are required in order to drive, maintain that automation strategy and initiative.
So, I think what’s really, really important here, automation isn’t just technology in isolation. It’s very, very important when you’re thinking about your automation strategy, you’re also considering the workforce transformation element, and that you’re also thinking about where you’re going to have to upskill and reskill your existing workforce, but also that you may have to recruit new workers in whatever shape or form that is, and also change those capabilities to adapt to the future.
OH: If you then apply an organisational lens to how you do automation well and at scale, you start to really challenge existing organisational constructs or structures, such as the relationship with the business and technology. Are robots technology and code, or are they actually digital workers and should be treated and trained like people? With so much of this new automation technology, it can be quite difficult and quite daunting for some companies to adopt it and to understand how they stitch different technologies together in the right way to work well with people.
It’s often much faster and more effective, actually, to move to outsourcing automation capability to providers of technology and skill automation.
Deloitte’s DARA platform and digital work is a great example of this. So, what that provides is digital FTEs using RPA or other intelligent automation technology, that can be purchased and trained so that they are stitched together to provide and solve different business problems. And that can all be provided at a fraction of the cost and without the burden of having to take on the technology infrastructure or the service and the controls wrapper that goes around that.
SS: And I think your answers there speak to something really nice about the two elements that you can get from robots. So, Oscar, you talked about replacing the human robot, which speaks to that consistency and productivity point of view. Gloria, some of what you talked around about the new skills, new jobs, and making the most of the human skills that we’ve got, speaks to how we start to unlock value. So, how do you find the right balance between that consistency and productivity point of view and the unlocking value?
OH: Great question, Sam. What we found is many organisations are doing proof of concept to have successfully automated areas, but then are finding barriers to really realise the full potential of automation and AI. Our 2019 global survey found that only 8% of organisations have actually successfully scaled automation. Although, that, we believe, is rapidly increasing.
And our survey which is out in a couple of months, we expect that figure to be much higher. What we do know is that when done well, there is huge value to be unlocked through automation and far beyond productivity.
GV: Around that, I think there are a couple of points I want to make in regards to this question. The first one is that automation doesn’t happen in a vacuum, right? Of course, initially, the motivation to start on an automation journey might be to increase productivity and achieve consistency. However, it’s much broader than that.
I’d say that beyond the element of more control, tighter quality and speed, there can also be a real impact beyond productivity. I actually wanted to talk about an example of an AI model that Deloitte delivered, that did intelligent triage at a hospital.
So, Deloitte worked with the gastroenterology department in a large hospital in the UK, and they basically developed an AI solution that improved the triage of general practitioner referrals. So, really, the aim was to use AI to unlock the data held in electronic medical records, and it allowed for more efficient processing, intelligent analysis, and improved decision-making in order to overcome some of the service challenges that they had been experiencing.
So, what was used was a combination of natural language processing that read the incoming GP referrals into the gastroenterology service, and it did two things.
Firstly, it suggested the most likely triage outcome, and then the urgency status and clinic or diagnostics for referral. So, this was all put together by RPA solutions as well. So, it was combination of the brains with the natural language processing model and the arms and legs with the RPA.
Now, the impact of this was huge. It was hugely positive. There was a 15% increase in speed, and a 15% to 20% reduction in the number of patients added to the waiting list. So, I think this shows that beyond productivity, there’s a massive potential for unlocking value in this particular example directly, societal and positive value for those patients, and the treatment of people who are suffering from cancer potentially. So, it’s definitely more than just consistency and productivity, I’d say.
OH: It’s a great example, Gloria. That one, essentially, AI enabled better decision-making to identify patients who possibly had cancer earlier, so they got faster triage, and they were seen by specialists. So, that’s potentially lifesaving. I think when you look at getting value out of automation, AI fails when it’s tech for tech sake. Automation should be rooted in driving up business value by understanding what the business needs and aligning to that, how AI and automation can best support rather than the other way around.
Gloria, perhaps, if I could bring you in here just to talk about how we see value being achieved through automation, beyond pure productivity, and also how we can use that to move beyond task-based automation to get much more value through hyper-automation and end-to-end automation.
GV: Thank you, Oscar. Sure. I think what’s really important here is to be clear from the beginning what value means for a specific organisation. At Deloitte, we look at value across key value levers of cost, speed, quality, and experience. However, each organisation has their own mission and vision statement, and therefore, their automation efforts need to go hand-in-hand with what value they want to achieve at the end.
Traditionally, when we look to get higher value, it’s not just focusing on task-based automation, or it’s not just focusing on automating [unclear] necessarily. To truly unlock the most value, you should really look to reengineer processes, improve how teams work, improve how processes are run, and not just look at the small vision of this specific bit of a process, but rather look at the end-to-end. That’s where you can achieve the most value. Really looking at the different hands-off, the start-to-end, even if it crosses multiple functions.
So, I think there are definitely ways in which you can unlock value practically when automating and when looking to incorporate intelligent automation.
OH: Thanks, Gloria. I think, practically speaking, cost is often the main driver of a business case for automation. But in my experience, when you talk to execs, they don’t just want cost out from a business. I think there’s a great opportunity, if you like, when you got the bonnet up on something, and you’re automating to taking cost out, to achieve those other value levers.
So, to make the processes more simple. To eliminate the process variations or the bottlenecks. And that’s where we really see value being maximised.
A good example of this is we automated an end-to-end complaints process at a major bank. What we did there is we started off with a manual process, but we implemented a small number of point automations using bots to get some quick wins and prove the technology works. We then moved into phase two, where we actually… The underlying platforms were updated, and we put in workflow to get much better end-to-end control, and we also put in further automation.
At stage three, actually, we then used the power of the data of those bots. We put it in an analytics platform, which actually took 1,300 variables and used those to identify those complaints which were actually at high risk. And this insight using the data that was provided from that automation, that technology, gave much better insight and reporting to the business. It also was much better for customers because they can self-serve and understand where their complaint was in the journey.
So, it’s just a great example of, yes, automation takes cost out, but those other value drivers really come in when you start to automate more of an end-to-end process.
SS: And the incorporation of automation into the workflow presents, yes, a really interesting question, especially linking back to what you were both saying before about how robots and AI need to become part of the workforce. So, quite simply, how do you line manage a robot?
GV: That’s a really good question and also a bit of a playful one as well. I’d say, initially, it all really depends on the type of robot, right? We’ve been talking about all these different technologies. We’ve been talking more about RPA, which is more task-based, rule-based form of automation, or the way to AI that kind of imitates a human reading vision and enables better decision-making.
However, a point I’d like to make is that I think a robot should be seen as something that augments the human and is more likely to augment the human worker to do a more diverse range of work than it is to line manage. We actually have a term for this within Deloitte. We call it the Age of With. So, human with machine. That illustrates that elements of augmenting and being able to work in a more efficient but also different way.
As I was saying, it depends on the type of robot also. So, RPA does exactly as it’s told. It’s black or white. However, when we’re talking more about our data model and something more on the AI space, that is something that learns. It requires data. It requires information, and then things become a bit more complicated because it sort of evolves with time and changes as you feed it different information. So, just like people, models need to be upskilled. So, for a natural language processing model, you’d need to feed it data sets so that it would learn.
So, really, depending on what you feed it, it will learn different things. But I think it’s interesting to see how the different technologies can help augment humans in different ways.
OH: Thanks, Gloria. I do think as organisations scale automation, there are some real questions there you need to answer. Essentially, yes, you do need to design and set up how to orchestrate and run automations, and how you drive continuous improvement from that workforce. And we see automation as digital workers rather than technology and need to be managed in that context to maximise the return on investment and utility you get from them, but also to keep upskilling them.
So, someone, yes, does have to have responsibility for orchestrating the robot workforce. Typically, this does require some form of centre of expertise, or excellence, or linked COEs, and a command and control or orchestration structure, plus some form of analytics to be able to really understand the data that the robots or the automations are providing, and then to drive end-to-end process performance and improvement.
In my experience, it’s quite common to have robots idle and utilised, which clearly isn’t what you want if you have a human or a digital workforce. I worked with… One organisation has some bots down there only working for 5% to 10% of the time. Imagine if one of our people was only working 5% to 10% of the time. There’s huge opportunity there.
Clearly, it’s not the bot’s fault. This comes down to how they’re managed and orchestrated. And when you start to run thousands of bots instead of ones or tens, this becomes really important to be able to effectively get the most out of your investment.
Solving this is really a combination of developing the right automation capabilities, and then keeping them upskilled in the same way as you would for a human workforce. And our Deloitte DARA platform here has in-built orchestration across different automation technologies to allow you to then orchestrate those multiple capabilities and skills to meet your needs, as well as some form of continuous upskilling in the same way you would continue to upskill your human workforce.
SS: And from a question that looks maybe slightly further down the road to one that’s being asked seemingly all the time, it’s in the press, and it’s been asked all the way back since machines started being brought in during the Industrial Revolution, of whether all this automation is a good thing.
OH: We’d probably debate for hours, Sam, on the merits or the drawbacks, and I think there are both. Whether or not people like it, I think automation is inevitable. In my experience, it can completely transform and be a very, very good thing, but I do think there’s a vital element here, and that’s the people element. Doing automation well has to be designed so that people and machines work together in the right way.
GV: I’d like to play a bit of a devil’s advocate with this question, actually. I think there’s this perception that when automation is used in daily life or when we encounter it in our daily lives outside of work, everyone embraces it and it’s a positive thing. However, it seems almost often when we talk about it in relation to automation at work, it’s seen in sort of more negative terms.
What I’d like to say is automation is already all around us. All of the apps that we use, they’re AI enabled. A lot of them, they have automation in them essentially. Your dating apps, your transport apps, everything, and even your banks. A lot of the apps are enabled through automation. And that seems to be perceived in a positive way. Whereas, in relation to this question, it’s almost like when we use automation at work, it’s sometimes seen as something negative.
I think change is scary. Potentially, if you don’t communicate well with your workforce and you’re not very clear with your ambitions and your strategy around why you’re implementing automation in the organisation, it can have negative connotations. However, it’s really important to be clear from the beginning what your strategy is, and what your choices are going to be, and why you’re pursuing automation as an organisation, and communicate that really clearly to engage the workforce and mitigate any adverse reactions to it.
OH: I think those who embrace it in the right way will get the most out of it. One of the banks I was working with created this concept called automation festivals. So, we ran these festivals in the UK and in India, which touched thousands of colleagues, and those festivals were a celebration of automation.
So, a combination of learning. You come to the work. You see it, and you can interact with it. We have different vendors there. But it creates a level of excitement and a buzz within the workplace that’s persisted around automation, and a workforce that really wants to get involved in it.
I do think, with all this automation, is it a good thing? There are lots of unanswered questions that we will need to address as a society. For example, we don’t want to introduce bias into AI through coding that may unintentionally discriminate against a particular gender, ethnicity, or certain demographics. Also, on a societal level, we need to think about fundamental reskilling of a workforce, so that those skills or roles that are automated, those people are trained up so that their skills are relevant and they can be productive in a future workforce.
SS: So, throughout what we’ve been talking about today, we’ve really brought out the people elements. It’s come up again and again. And the potential paradox has raised the idea of resilience, and I think that’s been brought out by the COVID-19 pandemic and the response to that. So, how can people and organisations grow to adapt? And what does it mean to be resilient in this world of AI and of huge changes?
GV: Thanks, Sam. I think this is a really, really good question, and really, really timely, actually. Linking back to what you’ve just said, the pandemic, the COVID-19 pandemic has really shown that if you need to change, you can really rapidly do so and really adapt to disruptive events quite quickly.
So, an example of this is the huge numbers of people that, within a day, went from working in an office environment to virtually, remotely working from home and actually thriving whilst doing so.
If I look at that sort of resilience aspect, I think resilience at an organisational level is something that includes your employee and your technology. So, it’s the ability to thrive and adapt in difficult times, or in times of massive change, but really having that perfect marriage between tech and people that allows you to do so.
OH: The FS industry is a great example here, one that traditionally had a bit of a bad rep in the past, but they’re using automation to be resilient and to make life better for customers. So, look at NatWest’s strategy, for example, of how they’re putting customer first in times of need, including loan holidays and mortgage payment holidays and helping customers with debt.
They’re using automation at scale to process huge amounts of volumes for customers quickly so they can be really, really responsive and help customers in times of need.
GV: Yes. And I think, as organisations move towards hyper-automation, this will require a fundamental reskilling. This will need to happen for the organisation to thrive. And so, as more sophisticated technologies start to grow and are more readily available and adopted, you’ll need new org structures and new skills to go with that, and to ensure that the organisation can be resilient to that sort of technological disruption, right?
I also think that getting involved in AI and experimenting or taking AI to whatever the next level is for each organisation is a way to predict and develop the skills that will be more relevant in the future.
So, I think automation is creating a reliance on reskilling, and I think a lot of the workforce will need to be resilient in this change as well. And as the work people do become materially different, and the way we work really changes, or the way work gets done becomes really different, it’s interesting to think about that resilience in light of technological disruption, and how to mitigate that and make a really positive impact with it.
OH: I think we should all encourage organisations and society to invest in reskilling to prepare for tomorrow’s world. I mean, if you look at the skills needed in the workplace of tomorrow, there’s much more people-based skills. Communication skills. Judgment and problem-solving abilities. Plus, also skills to work with technology. So, data science and analytical capabilities.
And on a personal level, I think we should encourage people to be aware of this change. The best way to adapt to be resilient is to learn about it, embrace it and become a part of it where you can. You might start off with something simple like process automation, RPA, but that can quickly mature. So, you up the value curve and reskilling as you move through cognitive and AI technologies.
SS: Thanks very much. That’s all we’ve got time for today, ending with a nice note to our listeners. Oscar and Gloria, thanks very much for joining us. And thank you for listening wherever you may be. Join us next time where we’ll be talking about workforce strategies and ethics as part of the perspective paradox.