RPA for Internal Audit, with Abigail Zhang, Andrea Rozario, Prof. Miklos Vasarhelyi — Webcast Q&A Transcript & Video

Edward Brooks: Thank you for turning up to this session, you’re going to hear from three experts in this space. I will allow them to do introductions in a second or two. Abby I think the background was that you and I had both worked on content at the Association of RPA last summer.

I think we’re both coming out from different angles and the more I was talking with you and your colleagues about what you were doing, it just seemed like a fascinating topic.

And interestingly one that we were getting a lot of demand for on the training side was what are the implications for RPA and robots in the audit space?

So suddenly there are opportunities and challenges that I saw opening up, so I really appreciate you taking the time to do this. Do you want to give us a bit of background? Just on how you got involved, what your history is and what led you to this place?

Abigail Zhang: Hi, we’re from Rutgers Business School, Continuous Audit and Reporting Lab. Today we are going to present our work related to RPA in in the audit area. Professor Miklos Vasarhelyi, myself, Abigail Zhang and Andrea Rozario will do the presentation together.

Andrea will talk about the first part of the presentation and I will take over the rest. Andrea and I are both PhD students at Rutgers Business School and Professor Vasarhelyi is a KPMG distinguished professor in the AIS domain.

Edward Brooks: Thanks, when did you get involved in RPA? What was the trigger point?

Abigail Zhang : We basically started a couple of RPA projects, about one and a half years ago and the motivation is that RPA has been used in the finance domain, in the accounting domain, but in the auditing area we still see slow progress, that’s why we started to look at the implication of RPA in the audit domain.

We basically cooperated with several medium-sized CPA firms and worked with them to see the potential of RPA in the audit area.

Edward Brooks: Just before you jump in, what were the barriers you received, why do you think it hasn’t been adopted more quickly in your field?

Abigail Zhang: You mean why RPA is slow in the audit domain?

Edward Brooks: Yeah.

Abigail Zhang: One of our colleagues, he went to interview the big four accounting firms and asked why RPA adoption has been slow in the Audit area. The common answer that they received is that the auditing profession is a highly regulated profession, so any progress in terms of technology tends to be slow because of all the regulations faced with auditors.

So most of the RPA projects are still in the pilot stage, however in other areas, for example Tax and Advisory, RPA is; I cannot say mature but it’s developing faster than what’s happening in the Audit area.

So that’s one side of the story, the other side of the story is that, since more and more firms are using RPA the auditors need to be aware of this condition because maybe there could be extra risks introduced by RPA.

However, most of the auditors haven’t realized this, that’s also the current situation of this whole RPA domain in Audit.

Edward Brooks: Excellent, I think it is interesting because a lot of the challenges around auditing bots are kind of the same challenges as auditing other things but people still go through the discipline of working out how to apply traditional audit processes in an automated environment or a bot driven environment.

I’ll let you go on.

Abigail Zhang: I think Andrea… you can start the first part.

Andrea Rozario: Okay sure, also just to add on to what Abby was talking about in relation to why it may be a little bit slower to adopt RPA in the auditing industry — we’re seeing that even for public accounting firms that are doing pilot tests for RPA; they’re doing this for non-public clients.

This goes back to what Abby said about auditing being highly regulated. I’ll give a little bit more insight into where accounting firms are at this moment with applying RPA technology. We know about Robotic Process Automation… you all probably have heard about this new technology.

Essentially why Robotic Process Automation is so prevalent across various industries right now is because it has the potential to (from a more integrated process) automate procedures or process activities that are very repetitive and that are very structured.

So what’s different about RPA, when you compare it to traditional automation, is that RPA is software that it’s very easy to use. If you want to program a bot to go to a website and fetch some information, it should take you less time than if you were to program it using Python.

That’s one of the reasons as to why we here in the CAR lab are seeing that professionals in different industries are using RPA to automate.

Abby and I are briefly going to cover the topics of how we can audit the clients RPA and also auditing with RPA using RPA as an audit data analytic tool.

Auditing a client’s RPA is interesting because it essentially creates a new type of IT audit. We have to… as auditors & internal auditors, audit the system, the ERP that the client has but now with these RPA bots, we have to audit the bots, to make sure that they are programmed correctly.

Also to make sure that the expected output that the bots are supposed to produce, is really what they’re producing, which goes back to the programming of the bots.

These are just some RPA statistics; Statista estimated that this industry is going to grow tremendously to 3.1 billion by 2019 and 4.9 billion by 2020.

You see that there is a need to audit your clients RPA because we’re seeing a lot of RPA adoption, not just in the finance or retail industries but across a variety of industries.

As for risks of RPA programs; RPA can obviously reduce the risk of human error in transposing the data from one source to another source, but it does introduce new risks. These are something that internal auditors have to be cognizant of.

First of all thinking about what these new risks are and also how they address those risks to make sure that these RPA bots are compliant with rules and regulations. These are some examples of risk in RPA programs… there could be inappropriate access to the RPA programs.

This gets to the concept of segregation of duties, like who has access to these RPA bots and are those the right individuals to have access to these bots. Also from an RPA bot perspective, you also have to think about the concept of SOD, i.e. Segregation of Duties.

With RPA you can automate potentially large parts of a process. Assume that you automate about 80% of a process, you can do that, however you have to ask the question “If we use this one bot to perform 80% of this process, are we as a company still compliant with Sox regulations?”

What we are seeing in practice is that although the one bot for RPA can be used to automate the majority of the process, there’s actually several different bots that have different responsibilities and this is done to make sure that these bots are in compliance with Sox because we don’t want to violate SOD even with RPA.

This is something that would be important to consider. There is also the issue of data security, we need to make sure that RPA has access to the right data and that when it does produce the results for performing specific tasks, that it goes to the appropriate individuals.

There is also the risk of weak maintenance of bots as well as other risks.

Edward Brooks: Can I just ask a question on that last slide?

Abigail Zhang: Sure.

Edward Brooks: Has any research been done on these potential risks? Has there been any research done to actually say ‘well actually the majority of risks are in the data security side or in the bad design’

Have you actually seen anything trying to look at what is really happening & where the real risks are?

Andrea Rozario: This is still an emerging area of research. Where we are from a research perspective is seeing where RPA can be applied and from having conversations with some professionals in practice, we are seeing that inappropriate access to RPA programs is a real problem and the concept of Segregation of Duties.

This concept has been around since SOX was required. SOX compliance was required, but in light of these new technologies like RPA, we’re starting to see that there could be a need to rethink Segregation of Duties from an RPA perspective.

If you assume that we have a bot that is executing parts of the process appropriately and a human is not manipulating it, then we have to think about the fact of whether SOX is still relevant in some aspects.

Because if the human is not manipulating the bot and the bot is executing the activities as it’s expected to, then perhaps we need to rethink this concept. So that’s where we are from a research perspective.

Edward Brooks: Okay, excellent!

Andrea Rozario: Okay, the goal of auditing RPA would be to make sure that in the current state of things and legislation, that these bots are designed to be compliant with SOX regulations as well as other regulations and to essentially make sure that the business logic that is embedded in these bots is correct.

We also want to make sure from a data perspective, that RPA is using the data that it’s supposed to be using and that it’s disseminating this data to the appropriate individuals or the appropriate parties.

Some ideas on how auditors can audit the RPA; well, first we need to understand these programs; we need to conduct some walkthroughs. We can request documentation from the clients starting from how many bots they are using in which part of the process and in which process activities are these bots performing and how these bots are maintained.

You also want to make sure that you observe the process that is being performed by bots and also talk to the RPA users to understand how these bots are being used. From your observation you will gather if these bots are being used like they are supposed to be used.

That’s the perspective of auditing a client’s RPA. Now moving on to a different perspective of using this RPA technology to perform better audits and to leverage this as a tool to perform auditing… here we have a framework that can guide auditors to use RPA in their audits. Auditing lags behind in terms of innovation, but RPA is actually very, very good technology to apply auditing to because it is the least disruptive by far.

That’s why it’s more feasible from our perspective to apply RPA to auditing. Some of the things that auditors may want to think about when considering RPA usage in their audits is what process they want to automate.

Practically any process can be automated but we really have to take into account what are the benefits of automating a process versus the cost. We want to think about what the actual output is and if the actual output is going to result in process improvements.

And then once we do that as auditors we have to identify the best candidates so that we can target one of those candidates.

So from an RPA perspective, ideally we want to target the low-hanging fruit and go for those tasks that are very, very structured but they’re so mundane, such as the process of collecting data. This is something that is fundamental to the audit but doesn’t require judgment and can easily be automated.

From the process identification phase, we can move on to understanding the process. This is where we would break down the process into very, very small pieces so that we can use those pieces and convert them to functions to embed into the RPA bot, because these little pieces are the steps that the RPA bot is going to perform.

In addition to that, this is something that doesn’t really get talked about very often but we think that it is essential to really be able to explore RPA to its full potential and that is the concept of audit data standardization.

If you train the RPA to copy & paste, the RPA will do that, but the RPA is going to do that following a very, very specific format. If you use the RPA on a file that’s formatted differently, it’s going to perform the task but it’s not going to give you the right results.

This is why it’s important to have an audit data standard. In this audit data standard you have labels that are named in a very consistent format. If you have your client that has five files, but those files have the same values but the labels are different, then in this audit data standard you want to make sure that the same label captures these values.

This is a way to format the data in a consistent manner so that the RPA can process the task that it is expected to process. Once we have the data in a consistent format and we understand the process from beginning to end, then we can move on to the prototyping of these audit apps.

So we can think of audit apps as RPA audit bots, this is the programming of these bots. Then we want to be able to measure implementation of RPA. We want to make sure that this RPA is resulting in high ROI.

We want to come up with some measures for efficiency, is the RPA really reducing the amount of time that it takes for this task to be executed and also from a quality perspective, is the RPA able to perform a better job? Is it able to detect errors when errors are present?

Edward Brooks: What about the quantity element as well, does it to do more sampling?

Andrea Rozario: Yeah, also like this is able to scale how much more work you can perform, that’s a good point, yes.

Edward Brooks: Okay perfect. The question is, you’ve done this, what is an auditor going to do if you’ve applied RPA to an audit process, what is it going to actually allow the auditor to do instead of the work that is previously done manually?

Andrea Rozario: Right… this is a very good point that you bring up. The whole point for having RPA to do this, is it’s obviously good to leverage technology but now for auditors the idea would be for them to focus their efforts on more analytical work and this is not just a case for RPA but in general. This will be the case for moving towards technology in the audit space.

So why have auditors perform very simple tasks like matching a data element from one file to a different file when the technology can easily do this as long as we can rely on the data? So that’s the whole point is that we have to start rethinking how we can do this work.

We think that essentially the work that auditors will have to perform would have to be more analytical and less structured, so moving more towards analytical work.

Edward Brooks: Does that mean fewer auditors in the Audit world?

Andrea Rozario: I wouldn’t say that means fewer, I think there will be some change in the workforce but I wouldn’t necessarily say that’s drastically eliminating auditors or that auditors won’t have jobs.

It’s just that the type of job that they perform is going to be different.

Edward Brooks: Yeah, I agree.

Andrea Rozario: Should I move on?

Edward Brooks: Yeah of course.

Andrea Rozario: All right, you can see here the different phases of the audit process. You can think of robotic process automation (which is here at the bottom) as the overlay software that connects the different phases of the audit process to form a systemic audit approach.

What we like to call it here in the CAR lab is ‘audit production line’. We have the bots perform the work, but this is able to achieve near end to end audit automation because going back to the point that we just made, we still need auditors to make evaluations about the results of these RPA tests.

Auditors are still going to be needed to come up with evaluations, for control risk and for control testing effectiveness.

Now we’re just going to go through a couple of examples as to how RPA could be used for auditing i.e. auditing using RPA.

In audit planning oftentimes we have to complete a checklist to begin a risk assessment process. This consists of checking the box to select the accounts that should be audited based on materiality or perhaps a different threshold.

How can RPA help here? RPA can open the folder where the financial statements or where the trial balance is located and can automatically open this file and copy/paste the revenue balance in this case.

We can copy the revenue balance and it can paste it into an audit work paper template, which calculates whether the current balance in revenue exceeds the materiality threshold and subsequently the RPA can take this information, so it can check the box on the planning worksheet.

Essentially the RPA can tell us as auditors, what accounts need to be audited; it can do that for us. So now I’m going to pass it on to Abby, she’s going to talk about applications of RPA for internal control testing and substantive testing.

Abigail Zhang: Thank you Andrea for talking about the first part; here I’m going to show some examples of using RPA in internal audit, internal control testing and the substantive procedures.

The first example is using RPA in revenue testing. Basically the idea is to use RPA to replace the auditor to perform the repetitive, mundane and rule-based tasks in revenue testing. For example, normally, the auditors will log into third party websites to get audit evidence, or the purchase order invoice shipping lists and will then do some basic comparisons.

We can just build a bot to do basically the same thing as that. The bot will just mimic the auditors behavior to log into the website and extract the data they need and perform the testing. The bot can be programmed to do, for example, the matching and the basic calculation.

In the example of the Automated Revenue three-way match, after the bot extracts the audit evidence, cleans up the data and does the basic matching… it can calculate the percentages of the discrepancies and compare the discrepancy rate with the materiality threshold.

The bot can also be programmed to generate an alert if the discrepancy rate exceeds the materiality. Basically if the underlying task is rule-based, mundane and repetitive, you can just program the bot to do whatever the auditor used to do. That’s one example in the substantive testing.

We also have one example in using RPA to perform the confirmation procedure. This is a real example done by Dr. Feiqi Huang who just graduated from our Business School.

He went to an audit firm and talked to them — they came up with a use case which can be automated by RPA and this use case is the cash confirmation process in which the auditors need to go to the confirmation.com to confirm the cash account balances.

The basic steps are, that the auditors’ first need to prepare a request form and then log in to confirmation.com to search for the client information one by one from the request form and then extract some data, do some analysis, download the corresponding documents and extract some information from those documents.

You can see that the steps are well defined and rule-based and it does not require the auditors’ professional judgment. That’s why we think this use case is a good one for RPA application.

Edward Brooks: I’m just thinking about the audit procedures generally. In many organizations the policies or the processes are kind of laid out, they’re reasonably documented to a certain extent when it comes to orders.

I think your department or your function in most organizations is probably an extreme end to having almost everything very, very highly documented before the work starts so you can create a very good starting position for that first part of the automation process.

Abigail Zhang: That’s a good point — you are right. For audit procedures, most of the work is well documented. This actually laid the foundation with audit automation, for example, in the planning use case that Andrea just presented.

Basically the auditors need to complete a document, which is called a planning worksheet, and it is in a structured format. It provides the potential for the task automation and also the automation can facilitate the documentation, because the bot can just grab the data and do the calculation and automatically fill in the planning worksheet by predefined rules.

Structured documentation lays the foundation for automation and automation also facilitates better documentation in the audit area.

Those are the four steps in this cash confirmation procedure and for the first step i.e. the preparation of the request form, Dr. Feiqi Huang did some process redesign here because as you can see from the screen on the left hand side, it’s the original request form used by the auditors which is in Word format. But in order for the bot to search on the website from the request form, it’s better to put the information in a spreadsheet format.

He did a process redesign to convert the original request form into this Excel type of request form. This is just for the convenience of the bot because the bot can just grab data from the Excel much easier than from the Word document — this is why he did the process redesign.

The process redesign is very important for our other RPA projects because sometimes it’s not so efficient to just automate the procedures as they are. Sometimes you need to do some process redesign to realize the full potential of automation.

After he did that process redesign, he basically just programmed the bot to do what the auditors used to do. He programmed the bot to open the web browser and log in to the confirmation.com website and then programmed the bot to extract the information from the redesign request form and basically follow procedures and predefined steps that the auditors do — he just replicated everything by using this bot.

Here’s how it looks like in the software. UiPath is just one RPA software that people can utilize — in his case, he used UiPath. The above two examples are how RPA can be used in internal control testing and also substantive testing. Then I will introduce how RPA can be used in the final step of the audit procedure which is completing an audit.

In an audit review procedure, the auditors always need to compare the data to the previous time period and the latest time period. Basically you can build a bot to replicate the auditors actions, to log into the third party website, download (for example, the sales listing) and then perform some comparison to see whether there’s any unusual fluctuation in the data.

The example could be; let’s say the auditors have two copies of revenue transactions, one is for the end of October and another is for the end of December. In this case, the auditors compare the changes in different time periods and see whether there are any abnormal or unusual fluctuations.

And if there are any abnormal fluctuations, the RPA can be programmed to send alerts to the auditors, after which the auditors will get involved in investigating further.

I just want to summarize by saying that many people ask me what I think is the best use case of RPA in the audit procedure. And my answer is that there’s no unique answer to this question because as long as the underlying task is repetitive, rule-based, standardized and of course has electronic data input, then the task can be a good candidate for RPA.

It doesn’t matter where this task sits. It doesn’t matter whether it is in audit planning or in the substantive testing, in the audio review or whatever — so long as the characteristic of the task meets the criteria, it can be a good use case for RPA.

To conclude we just covered two aspects of RPA in audit. The first aspect is that since our clients are using RPA in their operational procedures or accounting procedures, the internal auditors or external auditors should be aware of the risks introduced by RPA when used by clients.

That’s the first perspective and the second perspective is that we think the Audit profession can also utilize RPA as one of the audit analytics tools or automation tools to help auditors save the time they spend on doing the mundane and repetitive simple tasks and to focus their energy on the more complicated tasks that require their professional judgment.

That’s basically what our presentation is about.

Edward Brooks: The first question that came up was, ‘what are the barriers to auditors adopting this technology wholesale, quickly on that scale?’ Because to me, it seems like a no-brainer for auditors to be adopting this.

What’s are the obstacles that can hold up?

Abigail Zhang: You mean what are the obstacles of the prevalence of RPA in audit?

Edward Brooks: Exactly, yes.

Abigail Zhang: The first thing is that auditors themselves may have some concerns about using RPA which we already discussed, it’s mainly because of regulations. Especially for the public accounting firms, they are more concerned about this because whatever they do in audit will be reviewed or inspected by the PCAOB.

If the PCAOB does not have a clear or supportive tone in regards to using this kind of technology, the auditors will be very conservative about using it. I think that the main pushback is from the regulations.

Edward Brooks: What’s happening in the regulatory area? I’m sure they’re discussing bots, but is there an awareness of when it will be fully endorsed?

Abigail Zhang: Professor Miklos — do you want to add to this point because I think you are more clear on that?

Miklos Vasarhelyi: I am not sure of what to say about this, the PCAOB is still sitting around and worrying about sampling etc. How they should set up a task force, looking at analytic methods, they are very far and a little bit worried about Blockchain. They are very, very far from thinking about audit automation and the reframing of audit automation. I don’t think at this moment we can think too much about the PCAOB.

I think they’ll deal with it when there is complete data… when robots are doing every substantive amount of internal audit and service.

Edward Brooks: That’s interesting… but in the meantime is there anything to prevent individual public accounting bodies or internal audit functions within corporations? Is there anything preventing them from adopting these technologies?

Miklos Vasarhelyi: Let’s say the following; that’s my opinion not substantiated by facts, is that you’re going to find a lot of corporations. We talked about two of them actually — reasonably large and have staff that know RPA. They are picking up tasks that they think are labor intensive and creating little robots to do it without asking too much in internal audit.

We just saw a paper by David Wooderson (speaker at Brigham Young) where he inquired about the big four and what they are doing. Basically the finding was that they are using hundreds of bots, but they are not using a lot of bots in internal audit yet.

Edward Brooks: Do you see any industries or nations or countries that are further ahead in their adoption of bots for either side of the orders that we’re talking about?

Miklos Vasarhelyi: No I don’t think so, although there is a lot of voice about the Chinese being ahead of the US, I’d be very, very surprised if they are doing anything that we are not doing or even close to what’s being done in J&J, Siemens, and other companies that have ready started on this path.

It is very interesting because we talk to large companies and they are not surprised about robotic process automation, they’re asking when, how and how much does it cost. Then you talk about PCAOB or you talk to their audit team and they are on a totally different frequency. They are not even thinking about it too much. One of the firms seems to be doing a little bit more.

Edward Brooks: Do you see them joined up internationally as a cross firm initiative? Or do you see individual countries doing their own thing in the short term before they join up globally?

Miklos Vasarhelyi: I think that’s going to be very heterogeneous. I haven’t seen any of the firms do it, unless they are very secretive about having a major robotic process automation initiative in the short term.

I think that one of the problems is when they realize it doesn’t take a genius to do it — it really cuts the costs of performing audits.

Eventually we’re going to fees schema and they don’t particularly want to see fees being cut. I think that there is some hesitation there. I think what’s going to happen is, some smaller firm or firms that are not very big in a particular industry are going to say “I don’t have a lot to lose, I’m going to automate a lot here” and then do it very low for business.

Then the pressures are going to start to grow and I expect that they’ll start to grow fast.

Edward Brooks: One of the things you didn’t talk about that you had talked about in our preliminary conversations was, I think the phrase was “continuous audit” — do you want to just spend a couple minutes talking through that and how RPA can fit into that?

Miklos Vasarhelyi: I started this project on continuous audit at Bell Laboratories in the mid-80s and the first continuous audit application that we know of was the AT&T billing system. We published a paper in 1991 about it.

Since then both the AICPA and the IaaS and ISACA came out with some kind of guidance in this line.

Now what we predicted and what they actually developed there and then in a bank in Brazil — it sat around for a while but now two of the big four companies have a practice that they call ‘continuous auditing practice’ but they are not adopting continuous audit for obvious reasons.

In the external audit, they are really kind of consulting to the internal audit department to do a more frequent audit.

Several firms now say that they do continuous audit… we have a little bit of terminology disagreement with them. Doing an audit at the end of the year and then doing it ten times a year or five times a year it’s not a continuous audit.

What we see continuous audit being is a totally driven automation verification process that’s an overlay on top of regular processes. What we have seen in technological development and ubiquity of telecommunication (like the Internet) has created a whole new scenario of what can be done.

If you look at the dissertations of our PhD students, they are doing lot of predictive audit, machine learning etc. The RPA effort is very interesting because we were part of the AICPA audit data standard and now we’re basically arguing that RPAs efforts have to have a data standard to homogenize different data flows.

And if you homogenize data flows, there is no reason to do an audit or a verification process once a year; you can do it nearly constantly and monitor exceptions. Not a lot of conceptual things that are nice out of this because what you say, “Well that’s not an audit, that’s a major control” — that’s just a question of terminology.

Also it doesn’t fit very well in the traditional four stages of the audit. It actually works in an integrated way of risk assessment and at the same time, analytical substantive tests and etc.

But I think the ecosystem of data, interfaces of data standard, analytics running constantly, predictive analytics and basic interaction of data before the transaction, fall into downstream processes. It is something that is ready and part of the second system. There is going to be substantial automation using what we call RPA today and what we’ll call it in the future IPA and other forms of automation.

Edward Brooks: That’s fascinating; I guess it is kind of easy to pick up your passion but also the difference between where your mind is going and where your team’s research is going, where the kind of the real world is.

What I want you to close on, is for each of the three of you to give a prediction of the one thing you think will change in the next 12 to 18 months in this area?

Miklos Vasarhelyi: I think that there will be adoption of small RPA steps, I call it baby steps whereby certain things that will take a lot of labor will be automated and I think the pressure today is so auditors will come from seeing the internal audit using these tools and the clients think “Why the hell don’t you use this thing? It’s more consistent, it’s cheaper and you can do things that you can’t do manually”.

Edward Brooks: Abby, what’s your prediction for the next 12 or 18 months?

Abigail Zhang: I agree with Professor Miklos that more and more baby steps could be taken by firms or internal audit functions or external auditors. For longer term, I think there will be a shift in the types of work that auditors do because we see the trend where the basic and repetitive steps will be taken over by bots.

The auditors will be required to have higher-level skills, for example, the critical thinking and creativity or like whatever high level analytic skills that you can name, that’s my prediction.

Edward Brooks: Andrea, getting all these new technologies and given the historical skills of auditors, how do you see the other profession evolving to the right skills to actually apply these technologies?

Andrea Rozario: It’s a difficult question to answer but just to give you my opinion on that, there’s a lot of talk at least from what we’re hearing, that firms may be looking in terms of hiring. Some may want to hire scientists and then teach them auditing skills, others may want someone with auditing skills and then teach them the data science knowledge that they need.

I definitely think the auditing curriculum should change or I guess the accounting curriculum in general, but I do not necessarily think that accountants or soon to be accountants need to be experts in statistics or experts in Blockchain, RPA, etc.

What I do think is that we do need to emphasize more of those new technologies in the accounting curriculum and for these soon-to-be accounting or audit hires to have these very basic skills to be able to understand what the RPA technology is and how it could potentially be used in an audit and what are some of the risks that will need to be understood and evaluated for a client that uses RPA.

Also have some more applied knowledge, and for those, perhaps the case study methodology to be able to provide these students with the hands-on skills that they would need to not only perform audits of Blockchain or RPA, but also to perform audits using these technologies.

Edward Brooks: Excellent, I think we’ll all get back together in exactly 365 days to see what’s happened in the coming 12 months… but in the meantime, I really appreciate you taking the time to do this.

It’s a completely fascinating area and it has given real insight into how one profession is going to be radically changed by RPA and how it is also going to adapt to what it does to the functions that it works with.

Thank you very, very much.

Miklos Vasarhelyi: Just to add one point Edward, I think Abby can send you David Woods’ paper, he basically surveyed the big four and what they are doing and I think people that attended the seminar said they don’t have the paper, but I think he made it public.

You should be able to share them with anyone. It is a very interesting paper.

Edward Brooks: Fantastic! We will put that in the newsletter as well.

Abigail Zhang: Thank you so much, thank you for whoever is attending this webcast.

Edward Brooks: Thank you, excellent.

Miklos Vasarhelyi: Thanks guys.

Andrea Rozario: Bye, thank you.