PwC's Tech While You Trek

PwC's Tech While You Trek: Working Autonomy

PwC Season 1 Episode 21

Tune into another episode of Tech While You Trek to hear PwC Director Joseph Voyles talk about how working autonomy will unlock the immense value created when automation, robotics and intelligent systems operate in unison. Autonomous systems gather large amounts of previously inaccessible data from IoT sensors, transaction histories, machine data, human input and many other sources for analytics and feedback, resulting in better predictive maintenance and intelligence.

Tech While You Trek: Working Autonomy 
Guests:  Joe Voyles

Adam (00:08):
Hello, everyone, and welcome to another episode of PwC's Tech While You Trek. I am your host, Adam, and today we're here to talk about one of the emerging technology convergence themes, working autonomy with Joe Voyles. How are you doing, Joe?

Joe Voyles (00:19):
Great. How are you, Adam?

Adam (00:21):
I'm good. Thank you for coming by. Listen, why don't you start by telling us a little bit about yourself.

Joe Voyles (00:24):
I'm a director in the firm's emerging technology group. We have a small group of data scientists focused on AI innovation.

Adam (00:32):
We're talking to you, Joe, about working autonomy. Let's start by getting your best explanation of what that is.

Joe Voyles (00:38):
There is a lot of discussion in the industry around automation, but a lot of that automation focuses on automating kind of an individual task. When we think about working autonomy it's more about transformation of a process, so it's really about digitizing the process itself. So taking something that was completely manual and maybe even somewhat already digital and transforming it so that there's a place for all the activity associated with the process, the data associated with the process, the human interactions with the process, for it to reside in a way where everything's kind of recorded. I think you often associate technologies like AI and RPA with process automation, but I think the reality is that there are multiple tools involved when you talk about completely transforming how an organization conducts a function. AI and RPA are obviously critical components. You could also see IOT becoming a critical component as far as collection of information in the physical world that may inform the process. Robotics and drones, drones more so for information collection, maybe robotics more for actual execution of tasks, and even tools like Blockchain that can be used as kind of a system of record for that automation.

Adam (01:44):
So you have all of these systems working together. Talk about how that unlocks business value.

Joe Voyles (01:49):
The way of thinking about all these different components fitting together there are several layers, the first being more like a sensing layer. So this is where you might see AI solutions or IOT solutions in place to kind of collect and curate all the data in the world related to that process, and then AI comes into play in more of a thinking layer where it's being used for planning and decision making. And then at a certain point you need to take action within that process flow, and that's where solutions like RPA, maybe some robotics, maybe some drones come into play to actually take what the brain is telling you to do and act on it in the real world or in some other digital domain. Then finally, there's just this layer of governance that's associated with remembering what was done in creating this record of all the things that have taken place. And I think if you can do that successfully, each of those stages unlocks a little bit of value. So being able to collect all the information about your business and place it in a digital realm unlocks immediate value because you can now start analyzing that data, making decisions from that data, and even potentially using it to automate more routine task.

When you start moving into having an ability to think, now you're starting to replace some human judgment. Ultimately, you want to get to one decision for a situation so that the outcomes are predictable, and I think that's kind of a near-term need for a lot of applications of AI and automation, then as far as this layer of remembering or creating an audit trail that has some transparency that you may not have today within a manual process. So being able to track every single step of a material flowing through a process or a decision flowing through the process, that's what that's going to enable.

Adam (03:32):
How do you see companies beginning to adopt these systems in the present day?

Joe Voyles (03:37):
A lot of organizations are just solely focused on task automation that they're taking steps necessary to remove some of the human interventions from these processes that historically have just kind of emerged through chaos. So a need comes up, you put some people around it, they kind of operate this process, and tools and things like that are built around it, and it just serves as a business function. And I think today you're seeing a lot of clients realize that this process that had emerged through chaos could be better organized, so codifying business logic that maybe had resided in the heads of employees. So that's kind of the low-hanging fruit, and I think as people climb the maturity scale you're going to see more and more solutions that start incorporating more intelligent components, maybe being able to read documents and understand what's in a document, or be able to take in a video feed and understand what's going on in a video feed.

Adam (04:29):
They're watching videos?

Joe Voyles (04:31):
It's a very real use case. For banking branches, for instance, there's a process that those banks have to go through as far as the control, the deals with opening up the bank. So there have to be so many people involved with opening the safe, it needs to be open at a certain time, and there's procedures in place that need to be executed in sequence. And the way that that's audited today is a lot of times organizations just go back to their CCTV feeds and they review the cameras, but as you can imagine you can't possibly review all of that video footage for every single day to make sure that all these controls are in place, so how can we automate this by recognizing when these events occur within the video.

Adam (05:10):
For instance, someone walking over to the vault door, when that appears on camera an AI or robotic process could identify that and it could then verify independently completely without the need for a person that that had been an event that took place.

Joe Voyles (05:25):
That’s exactly the idea

Adam (05:27):
So what are some of the challenges that companies have found while trying to adapt this kind of stuff?

Joe Voyles (05:33):
They're just common concepts of paving the cow paths. So the idea behind that is that if you pave the way that cows move through the fields, it may not be paving the most efficient process for moving your cattle around and if companies take the same approach where they are simply going through and they're automating the things that exist as they exist today, that may not be the best long-term decision for that process. You may actually be adding some additional complexity down the road that will be difficult to unwind. So I think having an understanding of if you make a decision to automate a certain task or a certain process or transform that process, are you taking steps to redesign it and rethink that process rather than just simply overlaying these technologies onto already a broken system?

Adam (06:24):
How would you go about identifying an automation that was planned properly that was useful versus one that you're describing?

Joe Voyles (06:32):
It's really more about creating a big picture view and perspective and taking a step back and looking at the automated process as a whole. So I mean, if you came to me today and you said that I want to automate this function that an organization is performing and I went in and I interviewed everybody and figured out here are all the steps, here are all the systems, and I put that on paper, and it's just this gigantic mess of connections and a huge web that's hard to unwind. That's kind of an immediate indicator that we need to take a step back and make sure that this is the most efficient way to perform this function. So I think having some understanding of what the current state process is and looking at it from a perspective of redesign versus solution for automation is the first step and the first way that we would approach it.

Another significant challenge that we see with automation adoption is just once you've taken and you've automated the things that are easy to automate, that just connecting systems and creating information flow is something that doesn't really require a lot of complex code and you get into the domain of adopting artificial intelligence to help push the automation along or move up the complexity scale. Organizations often lack the data to do this, meaning they historically have not recorded the decisions that humans are making or how they're making those decisions in a way that you can build the models, and that is a significant investment if you don't already have that data and often a capital cost that organizations just aren't willing to take on. The issue arises when they aren't willing to make the investment and they try to force technology on the problem in a way that's not efficient and also not effective, and that actually gets at maybe a final issue that we see a lot. It seems like a no brainer, but I think a lot of organizations miss the opportunity to think through how would I measure the success of this automation and do I have appropriate metrics in place to tell me whether or not it's successful? So not only knowing what to measure, you have to measure it. And if you don't build in those measurements, it's very difficult to make the case that you've made a wise investment.

Adam (8:54):
All right. So before I let you get out of here, I get to ask you one kind of fun question I'm asking all of our podcast guests. Are you quite prepared?

Joe Voyles (9:01):
Yep.

Adam (9:02):
What would you of 20 years ago be most surprised that the you of today is using from a technological standpoint?

Joe Voyles (9:09):
I've always been someone that's kind of on the leading edge of technology and adopting things probably far too soon, so I'm a bit numb to technology now in that the things that I'm using today are really not that surprising. The one, though, that probably, and this is a technology that's been around forever, home 3D printing. It's still very nascent to some extent and really only for a hobbyist. But I do remember the early days of the internet and you had peer-to-peer systems floating around where it caught all of these industries off guard that their product could be just simply distributed freely among people, and you had all those videos about not stealing a car so why'd you download music. But a 3D printer that's technically going to be feasible at some point in the future, you would be able to download an entire physical object and manifest it in your home with these printers. So that one, I think, is probably kind of a real one that's still a bit surprising to me.

Adam (10:04):
Well, listen, Joe, thank you so much for taking the time to join us today.

Joe Voyles (10:07):
Thank you.

Adam (10:08):
So thank you for listening to another episode of Tech While You Trek. I have been your host, Adam, and tune in again in two weeks for a brand new episode.

Speaker 1 (10:19):
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.