In conversation with Deepak Seth, a global technology leader with 20+ years in IT strategy, innovation, and enterprise architecture. As a strategy and business transformation program leader with Xerox, Deepak brought innovative and disruptive technologies into the IT environment of the company, building the capability to innovate quickly with innovation hubs that could prove out technologies like Blockchain, AI, Microservices, IoT, and others. Recently he was at Accenture as a part of their CIO Advisory practice dealing with Innovation and Thought Leadership. In conversation with Sanjay Ghoshal on the future of AI, the role of humans in digital transformation, and the disruptive and transformative role behavior technology can play in workplaces of the future.
Sanjay: As a thought-leader in digital transformation, you often talk about the missing human catalyst in digital transformation. What are your thoughts on this?
Deepak: Humans need to be an integral part of the digital transformation or it will not drive value to the extent that it can. A better way would be to use the digital transformation to accentuate what humans do or to extend their capabilities. Because while machines and digital processes are more efficient and effective, they are not more intelligent yet. We are still far away from artificial general intelligence (or AGI) where artificial intelligence can really think.
This was seen when the pandemic happened, and datasets that were used to train AI algorithms in the retail space stopped working effectively. This was mainly because they had not factored in the kind of changes the pandemic brought about in terms of human behavior.
AI from that perspective is still a lot of ‘garbage in, garbage out’, so if the datasets are not good, the AI will be bad. The human ability to perceive and judge information, that is still very much needed.
The human ability to perceive and judge information, that is still very much needed.
Sanjay: Today we are hearing much about the VUCA world and so much is happening because of COVID-19. How do you think a company can make themselves resilient and relevant in this kind of a VUCA world?
Deepak: I had spoken about this a while ago. Resilience can be broken down into 4 distinct components – people resilience, process resilience, technological resilience, and information resilience. In each of these spaces, the behavioral aspect can play a role. When people consistently practice the right behaviors, they become more resilient, they become stronger, they learn to do things better and it reduces the uncertainty. An employee who exhibits behaviors that don’t align with the objectives of the company will drive the resilience down. But if all align with a common purpose, the resilience of the company and the individual increases. Similarly, in terms of the process, because it allows consistency of process, which results in better resilience.
Sanjay: Behavior Science has gone quite deep into digital companies – but most of this focuses on consumers. What role do you think behavior tech can play in helping employees become more productive, improve performance?
Deepak: It can play a very important transformative and disruptive role. The basic framework of Behavioral Economics with Richard Thaler and the Nudge Theory was driven by consumer behavior.
I have a personal experience with the nudge. I have an iWatch that has these rings that let me know how I’m doing. There’s an exercise ring, a move ring, a stand ring, and it makes me feel good when I can complete all those rings. If this messaging positive reinforces me as a consumer, or as a human being, why wouldn’t that work for an employee? A nudge could reinforce an employee who is exhibiting the right behaviors or is achieving the right targets or has to be prevented from doing wrong behaviors. Especially if this is delivered digitally through a smartphone or a wearable device. Definitely, a lot of scope for it being driven towards employee behavior.
A nudge could reinforce an employee who is exhibiting the right behaviors or is achieving the right targets or has to be prevented from doing wrong behaviors.
Sanjay: Today, managers and the hierarchy face a challenge of having too much information and too much analysis – how can digital tools help them become more effective?
Deepak: Given this plethora of information, I think what managers need is drivers for action. They need to know what they need to act upon, moving from business intelligence to actionable intelligence and to the next stage where technology helps them drive the action. That’s where behavior-led technologies can play a role.
From a visual dimension, we used to have heat maps. If those heat maps could be accompanied by a virtual prompt, from a watch or a cellphone – that would work. A notification, an alert, that kind of nudge can play an increasing role in driving more productive behaviors.
Sanjay: For a frontline person, the right behaviors at work would be making the best practices a habit. What behaviors do you think managers need to practice for them to become more effective?
Deepak: Interactions basically boil down to people, processes, technology, and information. To a large extent, managers deal with people. Often managers may not be exhibiting the behaviors that make them more effective in dealing with people. For example, the right behavior for a manager could be to praise the employee when they do something good. Now, if that is a behavior that needs to be encouraged, you need a mechanism that gives the manager feedback. Like saying, “Your praise-o-meter is showing that you have not praised your employee enough today or not celebrated success with them enough today.” That could be positive feedback for them.
Right now, what happens is people get bogged down with the recency bias. So, if at 4:00 p.m. I praise someone, I’ll think that the whole day I praised everybody a lot, or exhibited the right behaviors. But if all the information is available, it is mined, which is a digital solution, followed by sentiment analysis of all the content, it can give feedback on whether overall the sentiment aligned with what needed to be exhibited or not.
Sanjay: Frontline employees get overwhelmed with the amount of digital exposure they have and the amount of information thrown out to them. It also impacts their productivity and motivation. How do you think leaders/managers can help the teams deal with this?
Deepak: I think they can do a lot. Here’s an example from my past, one that might be familiar for anyone who grew up in India. When we were kids and needed to get a new shirt, our parents used to buy the fabric and get it tailored (a bespoke shirt). The tailor used to measure different dimensions before he or she made the shirt. But then scientists were able to figure out that all those different measurements could be boiled down to a couple of metrics – like the collar size or the inner arm length. The data was then further organized into small, medium, large, and extra-large. And more or less, those shirts fit us. Statistics were applied to the measurement data, which then boiled down to 2-3 metrics.
A similar kind of exercise needs to happen with all the other information which we have. In some cases, the reduction has not yet happened. In many other areas of operations, where there’s too much information and data elements, people will be able to identify what is relevant and focus only on those things. And that’s where machine learning, deep learning, statistical techniques will come into play. Ultimately a lot of that information is just noise, we can do without it and still understand what’s going on.
Sanjay: Whenever we discuss digital transformation or AI, there is always a fear among many business leaders – will AI replace human beings?
Deepak: First of all, this fear of AI is not actually a fear of AI per se. It is the fear that humans have towards anything new or different, especially on the technological side. It has happened through the ages from when humans first discovered fire. When the loom was invented in England, there were a group of people termed “Luddites” who went around breaking the looms because they thought that looms would take away their jobs as weavers. Horse cart drivers thought their jobs would go away when cars arrived. Yes, the jobs did go away but society evolved new kinds of jobs and new kinds of roles.
In the same continuum, now AI has taken that place of a new technology that we don’t fully fathom, and we don’t know what it can do to us, so people are scared of it.
We may not be aware of what doors or avenues artificial intelligence may create right now. But that doesn’t mean that we should take the view that all of it will be bad.
Some of it is because people aren’t able to understand why AI did what it did. That’s why the whole drive now is towards explainable AI. AI that can provide some explanation of why it did what it did, in a language that is understandable by humans.
Sanjay: Behavior science says a big part of the human mind is governed by irrationality, like emotions. When AI still doesn’t have emotions, is it possible to imagine a future where AI + Humans will be stronger than either alone?
Deepak: I was in conversation with an AI expert recently. He made an interesting observation that ultimately emotions are also electrochemical impulses in the human brain. Anger, for instance, is generated from certain chemicals released in the body, certain electrical impulses are fired by the neurons. So, what we are describing as emotions can very well be reduced to some equations or some set of data at some point in time. That could happen.
Even right now, with behavior-led technology, some of the feedback or the iterative loop to drive emotions is already in place, it has not yet been fully put into effect as a consolidated whole. For example, if I’m carrying my smartphone, and I have my watch on or have some kind of wearable shirt with embedded sensors, it can judge certain things that happen before I go for a cup of tea or coffee. So, it figures out the conditions – my heartbeat, body chemistry, etc. Then what prevents the next step from happening – that the impulse is sent to me that I want to get a coffee – that feedback loop can be driven that way.
Moods are determined by the environment, like lighting – which can be controlled by smart devices now. So, for instance, maybe some conflict is happening in the house. Suddenly the mood lighting changes because the smart device has detected that it needs to bring calm into the house. Those kinds of technologies will come. All the information is there, all the devices are getting implanted in our lives slowly and steadily. Human + AI has already happened.
Sanjay: We’d love to hear about your five mantras of AI.
Deepak: I think the five mantras of AI are:
- Think Big in terms of your vision – what you want to do with AI.
- But you also have to Do Small – in terms of how you execute AI. You’ll have to pick what you want to do. Often organizations may not have the appetite to do a full-blown project.
- Think partners – because companies may not have the capability to execute on their own. You have to find the right partners to help you do it.
- Think People. Artificial intelligence is as much about people as it is about technology. So even as you’re thinking about AI projects, you have to think about the impact it will have on people and also how people can help you move ahead with what your AI objectives are.
- And the last one was, Think Evolution. AI is not a “point of time” implementation but a continuum. Plan for the long run.
Sanjay: Any concluding thoughts?
Deepak: Every piece of technology that has emerged over the years has helped us become better humans, or at least we have strived to use it in a way that helps us become better humans. I’m sure artificial intelligence will help us become better humans, and will expose newer dimensions of the human experience which we have not experienced so far.
When there were no cars, there was some dimension of speed and connectivity missing. Then when automobiles came about, we discovered that. That helped us get better. When airplanes came in, it made the world smaller. We may not be aware of what doors or avenues artificial intelligence may create right now. But that doesn’t mean that we should take the view that all of it will be bad. As an optimist, I believe most of it will be good, some of it will be bad. Every technology is a double-edged sword. But hopefully, ultimately, it will all work out for the betterment of society.
