Quality of Work Life: an evolving definition
In this interview, Oded Karev delivered his view on robotic process automation, AI, cognitive capabilities and the main transformations impacting the future of work.
Oded is a thought leader and requested keynote speaker in the world of robotic process automation. He is a seasoned professional in the strategy and operations domain serving as the global General Manager of NICE Advanced Process Automation Line of Business, which covers the full RPA suite there. He moved to this role after serving as NICE Director of Corporate Strategy leading some of the company’s key growth initiatives.
Prior to NICE Oded spent a long tenure with Accenture’s Strategy Consulting practice where he specialized in delivering multi-channel strategy, operating model design and digital transformation projects for large banking and telecom enterprises.
Could you tell us a bit about NICE and how you came to Robotic Process Automation?
NICE is a large portfolio company that has over one billion dollars of revenue, a stable business built around over 30 years traded on the New York Stock Exchange, NASDAQ and so on. NICE specializes in what we can call ‘WO’ which stands for Workforce Optimization. This domain is part of the broader CRM ecosystem, namely the CRM systems that manage customer relations. We manage employees within that customer service environment. Workforce optimization, workforce management, most of these topics focus on the workforce and customer service management. This is what we do with, for example, inContact as our call center solution.
As part of the evolution of our portfolio NICE acquired a company by the name of eGlue: in 2010. It is the “e” for electronics”, and the glue that sticks things together, because that company was doing exactly that. It was sticking together electronic pages, and other applications. Today we are calling it Desktop Automation. Unsurprisingly, their domain of expertise was contact centers. So eGlue was doing desktop automation for contact centers before it was called desktop automation, or robotic automation.
eGlue was founded in 2002, and a significant proportion of NICE’s automation team is still based on eGlue people who have stayed with us since the acquisition in 2010, so I have the privilege of working with people who have more than 17 years of experience in process optimization. eGlue was doing attended automation in the contact center environment, which truly augments the human workforce - man working with machine, automating specific tasks on the desktop of employees with a focus on a very complex customer service environment.
This was quite successful for NICE between 2010 and 2013. In 2013 and 2014, we started hearing from our customers that they had been offered similar products which could automate tasks, not on the desktop but on the server, to take the task, put it in a queue, process it, and thus execute end-to-end process automation.
So, we checked the requirements and we understood that except for the server side, namely the control room and queue management task prioritization, we already had all of the more complicated components, such as: the robot, how to automate, how to interact with other systems, how to maintain the logic, how to build logic in a design environment without coding, etc… These components, which are usually challenges for companies, who want to develop RPA solutions, were already an integral part of our product.
In 2014, we launched the unattended automation product, the robotic automation product, and we immediately had some great results with two very important big customers who were part of the product journey. Telefonica in Spain (you can see a video on our website that speaks about their journey with us) and the global retailer IKEA. These two were the very first customers to truly invest significantly in automation at scale with our RPA product. That explains about how we came to do RPA, with the two sides, attended and unattended, and with desktop analytics, by using robots which collect data from the desktop and know what to do with it.
Compared to pure RPA players, what would you say are your main differentiators?
Alright, so not necessarily in that order, but the first one that we already talked about is the fact that we originally came from attended automation world and moved into unattended automation. NICE was a precursor, and we are the ones who in fact invented the terminology of attended/unattended automation.
We think this is a very strong differentiator, because moving from attended to unattended automation is much easier than moving from unattended to attended. The attended environment is much more complicated, as employees usually open a few windows at the same time, so the technology needs to support proactive automation versus reactive automation.
In reactive automation we are in a situation where a task is being executed by a human or a Machine to a robot, and when a task is being received e.g. an incoming e-mail , or an invoice perhaps, it goes to an unattended robot in order to be processed. In attended robotic automation, the robot has the capability to monitor the screen in real time, to understand what the employee is doing in real time and react to that which requires real-time monitoring of the screen. In order to support this complexity, we have intelligent robotic software that very few other companies have.
More importantly, we have the capability to do this without impacting the performance of the employee’s workstation. You need to think about this technically. If you were to install something on the desktop, that would constantly slow down how the desktop works, no one would want that product.
We have to find a smart way to allow humans to work with machines. It would be completely useless if the employees just take their hands off and say “now the robot is working”. What you need to do, is enable the employee to work while the robot is working in the background. And we have overcome all these challenges.
So, to move from the complexities of the attended world into the stability of the unattended world ( where the employee is not interfering with the robot, everything is very secure, and where tasks can simply be sent to the robots, even if it is not in actual real time) is a simpler transition.
We therefore have a significant competitive advantage in the industry and we are the market leader in attended automation for customer centric activities. If you think about it, this is truly the augmented workforce.
We have to find a smart way to allow humans to work with machines
In your view, what will the Future of Work look like for the Augmented Workforce?
Just think of what we can do with technology today. Humans have the intent, and with technology we have almost become Superhuman. We can fly, just like Superman, not physically but with planes and technology. We can solve complex problems, and even improve our well-being, with technology.
But companies have not yet fully invested in improving the well-being of their employees. Take for example when employees are executing a large volume of boring and repetitive tasks, this can create irritation, decrease motivation and result in mistakes.
Enterprises have not done enough to provide their employees with the tools to augment their performance and encourage healthy levels of motivation and well-being.
That is why our vision at NICE is Automation for the People, automating business processes by using the combination of robotic flawlessness and human empathy. We build software for people to work with machines.
And we recently took this to the next level by personalizing the NICE desktop robot, by naming it NEVA (NICE Employee Virtual Attendant). This personification of the desktop robot, with a friendly interface, is designed to reduce automation anxiety among employees, by reassuring them that NEVA is there to assist them to perform better rather than to replace them.
The NICE robots are specifically designed, with the employee in mind, to use the desktop robot like an assistant, in order to enhance and boost their performance. We are positioning ourselves as enablers of the augmented workforce. And this is true for every industry.
Essentially, every time an employee is interacting with a customer while using a computer at the same time (with a wait period for the system to work), NICE can help.
So, would you say that you are you bringing the products that we see on the market that are for the public, like Alexa and Echo Dot, for the employees themselves?
Let’s open a chapter about that. Let’s put things into the right and logical order as this is an important question and demands clarification.
Yes, for that matter, Alexa, Dot, they are mainly conversational bots, now chatbots, also known as conversational bots, and RPA robots are two very very different pieces of technology and this is a very important clarification that needs to be made as there tends to be market confusion as both RPA bots and chatbots are referred to as bots. The best example I can give you is a pen and a paper. You need both of them to write. On the other hand you could take the pen and write on the wall, or on the table (without using paper).
You can take the paper and use engraving or other techniques but it makes more sense to use the pen. They truly complement one another but the two are completely different, you wouldn’t ask yourself if they are the same, and still people are making the same mistake about chatbots and RPA robots being the same thing when they are completely different both in their underlying technology but also in what they actually do, not only in how they do it.
If you think about what the conversational bot is doing, in the business environment, it is replacing employees. Instead of chatting with a human, the end customer will be chatting with a machine, however, when you talk to a human in the contact center, he/she would also close the loop and type data into the business tool, management system etc…. Like in a CRM application. It can also set up a meeting up for you. The chatbot cannot do that, it can simply have a conversation, but does not have the capabilities to execute various tasks in real-time. The RPA bot, on the other hand, has these capabilities, but it does not have conversational capabilities.
And if you want to really do a digital transformation, and truly use chatbots, you cannot just automate the conversation, and then send the data to a human who will then need to ask for assistance. Instead of writing complicated integrations between the chatbot and all the assistance, it makes more sense to integrate chatbots with RPA robots when pursuing digital transformation initiatives.
If you take a look at this from a consumer standpoint, the same challenge remains. Take Alexa for example, Amazon has introduced Alexa for Business. Anyone who wants can go in the cloud and build their chatbot with extra capabilities, and brand it accordingly. But it is still limited to a conversational bot. If you need a transaction to happen, you need a process bot to fulfill this aspect.
People are making the same mistake about chatbots and RPA robots being the same thing when they are completely different
What is your view on AI and cognitive capabilities and the maturity of these technologies?
Now there is a huge amount of hype these days around artificial intelligence. Just like there was once big hype around Big Data. As you know, and until recently when you didn't have big data you could not sell software. Eight years ago, for example, it was all about “real-time”. And today it is all about AI - if you don’t have AI in your products you cannot sell software.
But the way the RPA market is using AI is by way of integrations, for example by using APIs to connect into the AI ecosystems. This means RPA tools can easily integrate with chatbots, such as IBM Watson, Microsoft Azure, etc… Now NICE wanted to go beyond that strategy and truly embed AI capabilities into our product to truly have a robot that is using AI to automate processes, and to truly take process automation to the next level.
Let’s take a look at 3 examples of how this can be done:
The first one is what we call Automation Finder. Based on our desktop analytics capabilities, the Automation Finder has the native capabilities to monitor a desktop screen, as I mentioned earlier. Our robot monitors the desktop to identify the employee’s activities in real time, and also captures the data. We then use artificial intelligence (specifically machine learning) to identify specific routines or sequences accurately.
All of the collected data shows the repetitive tasks executed by employees on a routine basis, which show the best potential for automation. This is true digital automation discovery, where AI is used in the data analysis. Average clicks, actions, and movements on the desktop are analyzed with data mining and AI analysis in order to see what can automated, and different automation opportunities may also be identified.
Another example I can give you is how our robot analyzes images on screens. As you know, its common practice to use screen scraping technologies to recognize images, when the objects are not available. But this approach is not stable as any minor changes will break the process. NICE uses AI to teach our robots to become better at recognizing images by using a very unique AI image focusing algorithm, which we call shape analysis.
We take an image of the screen, we run the image focusing algorithm, which converts it into shapes and we treat it the same way a human would when understanding what the images are. So again, it is not using RPA to enable AI. It is using AI to enhance RPA to better connect systems.
The third example I can give you is what we call the Reading Robot. It is the addition of text analysis to a library in the robotic flow, to make it able to work with unstructured data. The idea is to try to make sense of unstructured data and obtain a structured output.
These 3 examples create a good picture of how NICE has integrated AI into our automation solution. We truly integrate AI to intelligently automate processes, which is a real differentiator for us.
Is this what NICE calls the Advanced Process Automation, or APA?
Yes, the reason why we decided to call our offering APA and not RPA is because we wanted to position process automation within a wider context. By bringing AI capabilities into process automation, we are truly enabling humans to work collaboratively with machines.
The pure RPA market is relatively easy and straight-forward, and today almost everything that can be automated, has been automated. In view of this, it is very difficult to differentiate the RPA capabilities of the products available in the market today. It would depend on the know-how.
In general, the RPA market is highly saturated right now, but the actual market, comprising of automation opportunities, is not, and this is where NICE is different. We have been embedding AI capabilities into RPA from the start, and as a result, we are leading the way with a more intelligent and sophisticated solution that drives real-time business process optimization.
Today almost everything that can be automated, has been automated
With that the APA you talked about, how do you see the future of work? How will employees be impacted and work with robots on a day to day basis?
So first of all, I think it will change dramatically. Even though we can already see the initial effects today, we believe that it will continue to change in ways which we cannot even imagine. With greater workplace diversity, where robots and humans work together collaboratively, the nature of work will be more value driven, and humans will be enabled to express more empathy. As such, new possibilities will emerge in terms of value creation and unprecedented levels of operational efficiency.
Here are a few examples:
The first example is from a process automation which we completed for one of our leading customers, a big logistics company which handles international shipping. The process starts off when a shipping request is initiated, and the robot uses OCR (Optical Character Recognition) technology to read the form.
The robot then collects all the relevant data and digitizes it. Once the data has been converted into a structured format, the robots can process it in the back-office system, which may include updating various applications with the data. Many times, there may be a need for an employee to validate the data which has been captured by the OCR. As such, we have created an additional step. In this instance the data will be captured by an unattended robot, which then sends it to a human employee for validation (which may include a comparison).
The human then sends it back to the unattended robot again, so that it can complete the process automation, by performing various different tasks. Now, from these two examples, you can see that the skill set of the employee in the second flow example has changed dramatically, even though the process flow is being executed in the same fashion. Previously the employee was required to manually type in all the data from the form into various applications, while they were manually typing in the data, the employee was able to verify and compare the data at the same time.
The employee just needs to know how to read, write and understand which data goes where, in order to execute this task. Sounds easy enough right?
In reality, the team had to be replaced since the employees, who were so accustomed to typing, did not really enjoy working in an environment where manual processing was no longer required. As such, they needed to hire a new team and went about targeting millennial gamers. During the interview process, they asked candidates “did you spend your high school time in your parents basement playing video games ?”, and if the answer was “Yes, I love video games,” then it was the exact profile which they needed.
The next step was to create competitions, along the lines of which person is creating the best matches, who is doing the matches the fastest, who is completing the greatest number of matches or even who is matching the most accurately. This is a great example of using gamification principles to motivate their new team. The key take out is that the skill set of the workforce needed to change, in order to fulfill these tasks, which resulted in increased efficiency and fewer people were needed to validate the data.
But above all, this example shows how the skill set of the workforce is going to change. I’m using this example in order to say that we don’t know what we don’t know, but we are on the verge of a technological revolution. The last big revolution was the industrial revolution and before the industrial revolution, the tailors came and objected to tailor manufacturing lines that would affect their work and they said think about how many millions of people will be out of work. But the big clothing retail industry we currently have would not have emerged without mass industrialization, and this is something no one predicted.
Now just think about the supply chain of the retail industry, you need the logistics, such as trucks to deliver the clothes to the factories. You also need shopping malls with employees in the shop. There is a greater ecosystem here, and the millions of people that work in the retail fashion industry today, in contrast to the forecast made by the millions of tailors that said ”oh my god I’m going to lose my job''. With employee capabilities being augmented by robots in organizations, people are going to find ways to create value that we simply cannot predict.
With employee capabilities being augmented by robots in organizations, people are going to find ways to create value that we simply cannot predict.
Last question before that before we come to these kind of matters, a lighter one, is there a favorite automation to you, one that is very dear to your heart?
OK so first of all, I love them all, they’re all my children, every time I see a new use case, I get very excited, every time I see a new organization using our robots in a very creative way, this is something that piques my curiosity very much, but if I have to talk about one use case in particular, I would choose the use case where a customer interacts with the organization.
We use the robot to aggregate data in real time from the employee’s desktop screen. We then take this real time data stream and send it to a real time exchange system. With our product and the real-time capacity that we offer to their system we have allowed them to receive accurate recommendations for interactions with the customer. And this helps the sales team with their work.
I love this use case because with our real-time products, everyone is benefiting. Not only does the employee now have the tools to do better, be more efficient but the customer is also getting a much more personalized and efficient offer. The technology enables the employee to do something that is truly relevant to the customer and answers their specific needs – all enabled by these real-time screen monitoring techniques.
The ability to do real-time screen monitoring, to make real-time decisions, and to automate processes with repetitive transactions, is something that a neither a machine nor a human could do independently of one another!
This is a beautiful use case. A real example of a human-machine collaboration and augmentation of potential. This shows us what is possible and how the real-time collaboration of machines and humans makes everyone better off.
The real-time collaboration of machines and humans makes everyone better off