Hermann Sidhu, CPA, global assurance digital leader at EY, walks us through EY’s exciting new project to use drones to help audit large warehouses and outdoor inventories. The goal of the project is for drones to autonomously scan bar codes, QR codes, and other labels and transmit that data to EY’s online auditing platform. It has the potential to make audits faster and more seamless for both auditors and clients.
What you’ll learn in this episode:
- How drones can make inventory counts faster, more accurate, and less invasive for clients.
- How the drones would “learn” where to go and what to scan.
- How auditors can troubleshoot and verify drones’ counts.
- The limitations and potential pitfalls of using drones in audits.
Play the episode below or read the transcript at the bottom of the page:
Courtney Vien: Hello, I’m Courtney Vien, a senior editor with The Journal of Accountancy, and this is The Journal of Accountancy Podcast.
Today I’m speaking with Hermann Sidhu, CPA, Global Assurance Digital Leader at EY, and he is going to tell us about an exciting project that EY has in the works involving drones. Specifically, they’re experimenting with drones, using them to help out with audits. The drones are making inventory counts and they’re using AI to send the data to EY’s digital auditing platform. It’s very cool and he’s going to tell us all about it. But first, please listen to a word from our sponsor.
Courtney Vien: Hello, Hermann.
Hermann Sidhu: Hi, Courtney.
Courtney Vien: How are you doing?
Hermann Sidhu: I’m doing great.
Courtney Vien: Great. So first, can you tell me a little bit about the drone project?
Hermann Sidhu: Sure. You know, we looked at drones potentially and started testing the use of drones for inventory observation.
As you would know and our listeners/viewers would know, a critical part of the audit is the inventory observation when our clients have stock. And we wanted to see if there could be a practical application of really enhancing quality of the audits using the new drone technology. And also from an efficiency perspective, there was an angle to use drones. So that really was the starting point of it.
You know, we were seeing a lot of developments in the market from a technology perspective that got us excited. And our vision really was – you know, it was almost a natural extension of our platform. So if you think of how we audit, we have already given our auditors a mobile device, where when they go to our clients and do the inventory observation, they have ability to document their test counts via mobile.
In the background we are now capturing more and more of our clients’ data from an analytics perspective, so we have the perpetual records on hand. And drone was the next evolution of that potentially, to say, ‘can we also enhance that experience by using drone technology to really enhance the quality, the volume of the observations, and create a better outcome for our auditors and our clients?’
Courtney Vien: So you would have drones flying around a warehouse, say, and scanning the QR codes or the labels that are way up high?
Hermann Sidhu: That’s correct. And you know, we looked at – as part of our proof of concept, two scenarios. One was within a warehouse and one would be outside.
We would audit, for example, the world’s largest retailers, big manufacturing companies, big pharma companies, who have permanent, massive warehouses. The ability to go and map the warehouse allows us then to have the drone fly autonomously without having a human auditor in the warehouse trying to fly a drone and being disruptive, potentially to the client, etcetera.
And similarly, you take that same concept outside. You know, we would have clients that have timber – you know, mining clients, cars, automobiles. Another opportunity for the drone – you know, once you have the area marked out, the drone can self-fly autonomously and be able to go scan barcodes, RFID codes, QR codes, OCR, etcetera.
Courtney Vien: And when the drones would scan a code, they would automatically send it to your auditing platform, that’s right?
Hermann Sidhu: That’s right. And that’s really the critical aspect for us, and that’s what we are working on now – you know, that integration back to our platform is critical for us, because we want to make that as seamless and as real-time as we possibly can.
Courtney Vien: So how do the drones know where these bar codes are? Would there be a human operator using a video camera or something like that?
Hermann Sidhu: I think that’s where the setup comes in. So there is a sort of first-time setup. Because every client’s warehouse is slightly different, right? And their bar coding technique is slightly different. And that’s why it sort of goes back to my earlier comment in terms of the size and scale of this opportunity, it cannot be every client, but for the large clients, you know, there’s a setup required. You have to teach the drone exactly what they’re looking for.
But then after that, you know, once that is established, it’s sort of on its own and it has the ability to go and scan it. And you know, we do this from an audit perspective, but at the same time it’s no different than our clients’- our customers’ companies outside are using for their warehouse management. So if you look at the press, you know, big organizations are moving to drones scanning their warehouses, and we’re utilizing similar technology that our customers are using. Obviously we would be doing it in an independent fashion.
Courtney Vien: Gotcha. So how would an auditor know if a drone had made a mistake or had say read a label incorrectly, had missed a label?
Hermann Sidhu: Yeah, ultimately the auditor has to set the parameters of what they want the drone to look at, which area of the warehouse they would scan, etcetera. It could be 100 percent scan where it goes and scans every label in the warehouse. It could be to go and target every fifth box.
Whatever the auditing technique or sampling technique the auditor wants to use, the drone would follow that technique. If it misses it, ultimately whatever the drone reads has to then be tied back to perpetual records. So if there’s a missed label, it’ll throw a flag up, because it’ll be different than what the perpetual records of the company that you’re ultimately reconciling back with. And that would really be the check.
The other sort of – you know, if for example the auditor says, “Go and scan every box in this particular part of the warehouse,” and the drone just isn’t able to do that – for example, RFID could be blocked, for example, not enabling the drone to be able to read that. I mean, those are real-life challenges that have to be worked through.
But that in some ways is no different than when the human auditor goes to the warehouse as well, right? Boxes could be everywhere. You know, back in the day when I was a staff , we’d go to all of these warehouses. Some of them are great and they’re very organized but you could always go to a warehouse that is not that organized. And those are frankly the limitations of the technology, similar to the extra work that the auditor has to do when they’re in person as well. Those have to be worked out. Ultimately, we can be only as good as how organized and how well-laid-out our clients’ warehouses are.
Courtney Vien: Sure. Would there be potential for fraud? Say somebody decided to put the wrong labels on certain boxes? Would there be a way to check?
Hermann Sidhu: Exactly, right? I just go back to, that potential exists. The same with a human.
When we send our auditors out, it’s not just reading the QR code on the box, right? They key is to open the box, evaluate what’s inside the box. And the auditor, per se, would not do it on every box, but you know, using their judgment and professional skepticism, they would go and perform that check. Similarly, with a drone, you would want to make sure that you are utilizing the same techniques, because to your point, just scanning the bar code, you don’t know what’s inside the box. So the drone has a camera functionality, imaging functionality. So the drone – you would just have the client open the boxes up and similar to the human eye, the drone would have an eye to go and see exactly what’s inside the box.
So that is definitely a limitation, where we need to make sure that we think through the auditing standards and the professional skepticism that auditors acquire, and judgment, and make sure that, you know, the drone has the ability to have all of those factors built into the program that we launch.
Courtney Vien: So how would using drones make an audit more efficient?
Hermann Sidhu: I think a multitude of factors. One is just the speed. I think it’s 21 times faster than a human. So that’s just the speed at which you can activate a drone is significant. I think it’s also – I always go back – you know, for our clients it’s less invasive. You know, you want to make it less invasive. The drone can go when the warehouse is, for example, shut down. You don’t need operators in the warehouse. So you can technically fly the drone any time of the day and pick the most least-invasive time for our clients to do the inventory observation. I think that’s a big benefit as well. And then from a quality perspective, you know, it’s just, how can you go? You can do more counts.
If you’re in a cycle counting environment you can potentially go and do more cycle counts in a less-invasive fashion, because you have the speed element, once again. So I think the efficiency comes out, but the mitigating factor, to be frank with you, is efficiency for – it has to be the right environment, the right-sized client, the right-sized warehouse for this to operate in today’s environment. Because the offset is the drones are still expensive. To train a drone is still expensive, to map out a warehouse is time-consuming. So you know, today’s technology is very much to me limited to these niche use cases, whereas tomorrow, as those costs come down, you can see a much more pervasive use.
Courtney Vien: Sure. How would the use of drones change an audit from the auditor’s perspective?
Hermann Sidhu: Typically all of these inventory observations happen right at year-end. In my experience, the first couple of years in the firm we were told, “Listen, don’t try to make too many plans on December 31st, New Year’s Eve, or January 1st,” because that’s where the big volume of year-end inventory observations were. So right off the bat, if you can have a drone – you know, if you use a drone versus sending a human, it changes our people’s experiences. Some inventory observations we do, frankly, are in situations or circumstances that are just not the most conducive. You know, you could be out in harsh weather conditions, out trying to count timber, coal, etcetera, etcetera. So a drone, I think, would really enable and enhance our people’s experience in those situations.
And I think there’s a cool factor with our people that’s critical. And then from an audit quality perspective is – you know, you can program the drone, you can enable the drone at any point in time. It’s less invasive for our client and can drive much, much higher volumes of observation that we typically would do, using traditional sampling techniques. So I think that’s critical. And also, from a consistency perspective and insight back to our clients – you know we are in a situation today, for our large clients you could have hundreds of auditors going out. Think of a very large retailer with stores all around the world. You could have thousands of auditors go out and do inventory observations at different times of the year at all of the different warehouses. Well, with a drone, you can drive a level of consistency, because you can then get feedback on each of those locations using what the drone has observed, plus the camera, what they viewed,
package that together and bring it back to the client in a much more sophisticated fashion that shows what the consistency, inconsistency is or what the opportunities are for improvement, etcetera. So the technology can also add a significant amount of value and has a significant potential to add value and add insight back to our clients in a much more effective fashion.
Courtney Vien: Yeah, it sounds really cool. Would the auditors actually be flying the drones, or would they have somebody doing that for them?
Hermann Sidhu: You know, the drones are autonomous, and that’s sort of been our focus, is that we ultimately want to start where the drones have the ability to fly autonomously. So where the human interaction is, is in the initial setup – you know, EY would control the initial setup. The drones still have limited battery life, for example. So you know, someone physically has to be there to swap the batteries out. Battery life is getting better.
You have to obviously service the drone and maintain the drone, etcetera. So the operation – so if you think of us an audit firm, we don’t have those capabilities, per se, of maintaining drones. As we scale this, we have to think through all of the logistics of what that entails. But really the thinking is to create autonomous drones. That’s where AI comes in. Because the next-generation drones that I’ve seen have the ability where they have the ability to learn sort of the happy path of what the warehouse is for them as they’re seeing obstacles, etcetera, and the ability to fly around and correct course. The drones are getting much more smarter as well.
But I guess to summarize, the point I’m trying to make is that our view is that very much the drones would be autonomous. We would not actually have to get our auditors licenses in various states and jurisdictions to go and be able to fly the drone.
Courtney Vien: You used a phrase, “happy path?”
Hermann Sidhu: Yeah, so a drone – you know, if you think about it, the drone – you have to set up, if it’s an outside warehouse, you know, what are the parameters of where the drone is going to go? Or if it’s in the warehouse, you have to teach the drone, right? What are the different bays? What are the turns, etcetera? And that’s where the mapping of the warehouse comes into play. And that’s the initial setup which has to be conducted in a warehouse. And that goes back to an earlier point. As you would imagine, you would only go – you know, think of a big retailer with a massive warehouse that has just trillions of products in it. It’s efficient to go and map that warehouse out and then have the ability to use a drone to fly autonomously. Whereas, if you have a small warehouse, etcetera, you know, with today’s costs of this stuff, it may not make sense to actually go and do it.
Courtney Vien: So I know this project is in kind of its experimental stages. You just filed for a proof of concept. How far away is it from being actually put into practice in the real world?
Hermann Sidhu: Our proof of concept was successful; we’ve completed it.
Now we move from proof of concept to pilot. You know, we’ve gone out and we have a list of clients that are in a variety of sectors, both for internal usage and external usage, who want to, are eager to participate in our pilot program with them. And we are in the process of getting that rolled out, so whether it’s a big – retail, manufacturing, construction, agriculture, you name it, we want to get a cross-section of those industries. And then I think the next – from there would be – you know, the scaling questions would come in. What are the hardware requirements? Environmental factors such as weather conditions, flight risk?
So my view would be, by 2019 we want to roll out pilots on a multitude of clients, see what our success is, see how the integration back with our core audit platform looks like, and then from there take it and see how we scale it further. The process from proof of concept next phase, later half of ’18, ’19, we want to finish our pilots, then assess a go or no-go decision on how much we scale it.
Courtney Vien: Is there anything special that auditors would have to know or do differently in order to work with these drones?
Hermann Sidhu: No, I think the auditors – you know, our young folks coming in, they’re excited about this. They are eager to experiment. You know, they are supportive. It’s exciting for us. So no, I think it’s – no, I think, especially for our auditors going out on December 31st in cold conditions and having to do some of these observations, this could potentially be a very positive development.
Courtney Vien: Yeah, they can just sit in a warm office and watch the video feed.
Hermann Sidhu: Right.
Courtney Vien: Well, thank you very much for speaking with us. So nice to have you, Hermann, and you have a great day.
Hermann Sidhu: Thank you very much. Really appreciate the opportunity.
Courtney Vien: Again, that was Hermann Sidhu of EY, telling us all about their new drone project, and this has been Courtney Vien for The Journal of Accountancy Podcast. Thank you for listening.