PwC's Tech While You Trek: Digital Reflection

Tune into another episode of Tech While You Trek to hear PwC Partner Anand Rao and Senior Manager Sindy Ma talk about an emerging technology convergence theme called digital reflection. Digital reflection is a virtual representation of complex and interdependent physical processes and interactions that can be tested against real-time scenarios in order to predict an immediate outcome or inform business decisions. Digital reflection can include emerging technologies such as Artificial Intelligence, Augmented Reality, Blockchain, Drones, IoT,  and Virtual Reality. 

Transcript

PwC’s Tech While You Trek:  Digital Reflection
Guests:  Anand Rao & Sindy Ma

Adam (00:08):
Hello everyone, and welcome back to another episode of PwC's Tech While You Trek. I'm your host, Adam. And today we're here to talk about another emerging technology convergence theme called digital reflection, with Sindy Ma and Anand Rao. Thanks for taking the time. Why don't we start off by having you all tell us a little bit about yourselves.

Anand Rao (00:23):
Hi everyone. It is Anand Rao. I'm the global AI lead for PwC, and I get involved in all innovation projects for our clients, as it relates to AI, machine learning, deep learning, simulation modeling. I've been in this area since the mid-'80s.

Sindy Ma (00:39):
Hi, I'm Sindy Ma, a simulation modeler and AI practitioner based in Chicago. I have been building various types of analytical models for eight years at PwC.

Adam (00:48):
Let's dive right in. Can you please give our listeners an overview of what digital reflection is and the technologies that make it possible?

Sindy Ma (00:55):
Digital reflection is all about creating virtual replicas of real world systems, processes, or people, so that we can have a safe testing space. So imagine practicing a golf swing at the driving range. We want to be able to try that thousands of times with minor adjustments, so we can thoroughly understand what makes something successful or not. We think of digital reflections is an umbrella term for digital twins, simulation models, and other computerized reflections of the real world.

Sindy Ma (01:21):
I'll give two examples. First one is ride sharing service is a digital reflection, which incorporates a lot of different technologies. You would use a driver's phone as a sensor that sends geo-locating data that can be visualized as a vehicle icon on a map. And then you would have to incorporate AI and simulations, in order to be matched to a driver and allocated an ideal path to get to your destination.

Sindy Ma (01:42):
We also build a dual reflections that are at a higher level of obstruction for more strategic problems. Some examples here might include simulations that are designed to test policies that address issues such as rising healthcare costs, or maybe encouraging the growth of a business.

Adam (01:57):
How would a virtual model go about testing policy? What data and what tools would have to be at its disposal, in order to get to a point where it was doing that?

Sindy Ma (02:06):
It's very important then for us to be able to replicate the real world, understanding why things happen. And then we can test a policy on something that hasn't been observed. So we have no empirical data about how the system might behave in the future. By using those causal links, we can make a good approximation and guess at what might happen.

Adam (02:25):
How is digital reflection different from digital twin, or is there even a difference?

Anand Rao (02:31):
The way in which digital twin has been used in the industry is very specifically to simulate or to create a digital replica of a physical asset. So common examples used are it's an aircraft engine sensor, autonomous way like Sindy was mentioning. So those are all digital twins where you have a replica of an asset. The reason why you have a replica of an asset is obviously you can't do much with the engine for the new aircraft that's flying [crosstalk 00:03:01] about, but you can do things with the digital replica. You can stress test it. You can put it through various measures to really road test how well it behaves.

Anand Rao (03:10):
So now we introduce the word digital reflection, as Sindy was saying, as a much broader term. It's not just physical assets that need to have a digital twin. You can think of digital twins for even consumers, for people like you and me.

Anand Rao (03:25):
Now, the reason why we want to do that is we want to simulate our behavior, what we sometimes call as behavioral simulation. Let me just take a very simple example. Now, if you tell me it's good for my health to have a controlled diet and exercise 30 minutes a day, I might say, "Yeah, it sounds like a good idea," but I may never end up doing. If you show me what will happen to me 10 years out, if I continued with the same dietary and physical behavior, and all the complications that I could have in life, if you actually showed the real me in 10 years into the future, I might think twice. And that's what the digital reflection is doing. It is taking me as of today with all of the different parameters, and then it is simulating me into the future, so I can see that future self now and then take the right decisions.

Adam (04:20):
So then why is digital reflection on the convergence themes list?

Anand Rao (04:24):
The way we have looked at convergence is anywhere where there are two or more essential technologies coming together, we call it a convergence theme. Digital reflection combines the AI and the IoT. So there are a number of sensors that are in industrial locations or anywhere, where we see in our house there are sensors attached to my temperature reading, even my bulbs, everything has sensors. Now that is IoT. So it is capturing a lot of information.

Anand Rao (04:56):
The digital reflection has a digital replica. So now when we use AI and IoT together, we can start tracking how warm my room needs to be when I'm here, what is the outside temperature. So all of those adjustments are being done. And now my sensors and the AI behind them are much smarter in terms of my ideal temperature. So it starts learning more that even I may not be able to articulate. So that's why we have included digital reflection, which is a combination of AI and IoT.

Adam (05:29):
So can you tell us please about some of the use cases you've seen?

Sindy Ma (05:33):
One fascinating example that I love talking about is when we model brand new markets that don't exist yet, for example, autonomous ride sharing, drone deliveries, and hopefully one day, personal robot assistants. There is no previous data set for us to draw wisdom from, for these. So we really need something like a digital reflection, in order to test hypotheses for how these things might work. So any company that's hoping to build these would want to know what exactly is feasible and what risks might be lurking around that they need to consider. And in these examples, the digital reflections really help us take what we do know. So we currently know things like the population density of a state or what their regulations are for drones, and combine this with unknown. So for example, we might guess at the speed of uptake or competition behavior, and simulate these out till we can divine a good policy to actually implement these in the market.

Adam (06:22):
How has adaptation evolved over the course of the COVID-19 pandemic, and how are companies may be using it differently now than they were pre-pandemic?

Anand Rao (06:30):
That's a great question. So anecdotally, we were seeing more of AI and specifically more of digital reflection since the pandemic hit, and everyone is going online, so AI is gaining more prominence. So we just completed a global Responsible AI survey of more than 1000 executives across US, UK, Japan, India, and what came out was very interesting. So the investments in AI have definitely increased in US and in some of the territories as well. And not only as the investment increased, what is also clear is companies which were more advanced in the digital revolution or adopted digital were doing better in the pandemic and investing more in AI.

Adam (07:16):
Well, okay. So then why is a simulation being used more?

Anand Rao (07:20):
Specifically, as it relates to COVID-19, what we have seen is there is more of a demand for simulation-based services. The reason for this, I think, is quite simple. Executives make decisions either based on their intuition or on data or some model. Now, intuition doesn't really work in the case of COVID-19 because we haven't seen one like this for the past 100 years. The data that we have doesn't really work as well, because before what happened in COVID-19 and post COVID-19, there's very little resemblance.

Anand Rao (07:56):
So what should happen is someone needs to take some of these previous instances of diseases, model them, model the behavior of people, what we have been observing so far, even if it is just two weeks, three weeks, four weeks, observe those behaviors and model them. And that allows companies to see how they should be prepared for maybe the second wave or the third wave of the COVID-19, and also what they need to do to get back to work or return to work. So that's why we are seeing more of the use of simulation and digital reflection in some of the work that we have been doing.

Adam (08:37):
So what are the challenges that you've discovered companies have found with adaptation of this technology?

Sindy Ma (08:43):
Companies, especially large established companies, have their data in all different systems and everything is quite siloed. So just getting one cohesive data model is in itself such a huge challenge, but a lot of these systems also often contain predictive capabilities and sometimes even autonomous. They can take actions by themselves when certain triggers are hit. And so often we'll see that adoption actually stops short of dynamic models, which are really necessary to describe real world causal links, such as feedback loops and delays. Without these, it is really difficult to predict important phenomena, such as network effects or vicious and virtuous cycles.

Sindy Ma (09:21):
There are many reasons why dynamic modeling capabilities like this are not as widespread or spoken about. First one is quite simple. It's hard to do. Often, good is good enough, and the issue that a more dynamic model would have predicted may only come to fruition three years later, beyond the current planning horizon.

Adam (09:38):
Well, so as this technology and as these systems move forward, where do you see it going in the next decade?

Anand Rao (09:46):
Yeah, the next decade, I think it's going to be very interesting. So we have seen a number of things happen in the software world, in the open source world, and obviously in the AI world. And where we are going with this is that I think there'll be more open source simulations. So just as we have open source software where everyone plugs in their little module and can build from other people's software, I think simulation will also be reaching that stage 10 years from now. So there will be open source simulation platforms, where someone might have a model or a simulation of the economy. And again, different, no one really agrees on how the economy works. So you'll have competing models, if you like, on the economy.

Anand Rao (10:32):
Now, if I'm a company coming in, introducing a new product, and my product is very much dependent on the interest rate of the economy, I should be able to plug in and evaluate my strategies across a wide variety of different economic situations, not built by one individual or one university or one company, but across a wide variety of them.

Anand Rao (10:55):
Then from there, I think we'll also be evolving into more simulation as a service, where just as we have the AI platforms of today offering analytics and prediction as a service, we'll start seeing simulation as a service as well. And I would really see, 10 years from now, we could basically see digital reflections of us. I see this more as we can do a time travel or a future travel into 10 years out and say, "Hey, how am I doing 10 years from now, based on my finances, my health, and so on?" Make those decisions, come back and say, "That was not that good. Maybe I should exercise more." So I see a really exciting future for digital reflection.

Adam (12:30):
So before I let you go, I have one fun question to ask that I ask all of the guests. Are you all quite prepared to answer this question?

Sindy Ma (12:38):
Yep, absolutely.

Anand Rao (12:40):
Of course.

Adam (12:41):
So the question is, what would the you have 20 years ago be most surprised, from a technology standpoint, that the you of today uses?

Anand Rao (12:50):
25 years back, I was building simulation models for top gun pilots to simulate air combat situation. So at that time, I don't think I would have thought, or I didn't think that I would be using more or less the same technology, 20 years from then to model human behavior in the businesses and all the healthcare financial decisions. I never thought it would be similar software that would be so popular in the business domain.

Sindy Ma (13:21):
I'm just really shocked at the faith that we place in our 5g, in our cellular networks. If that fell apart today, we wouldn't be able to do things like pay for anything or use maps when we travel. So that's probably what I'm most surprised about.

Adam (13:34):
Well, listen, thank you both so much for taking the time and stopping by the show today.

Sindy Ma (13:39):
Thank you.

Anand Rao (13:40):
Thank you, Adam.

Adam (13:40):
Well, thank you, Anand and Sindy for joining us today. We hope everyone learned something new about emerging technologies. I know I did. This has been another episode of Tech While You Trek. I have been your host, Adam. Thanks for listening and tune in again in two weeks for another new episode.

Speaker 4 (13:58):
This podcast is brought to you by PwC. All rights reserved. PwC refers to the US member firm or one of its subsidiaries or affiliates, and may sometimes refer to the PwC network. Each member firm is a separate legal entity. Please see www.pwc.com/structure for further details. This podcast is for general information purposes only and should not be used as a substitute for consultation with professional advisors.