Select Page
audit planning keyhole illustration
ON-DEMAND WEBINAR

Data-driven audit planning: The key to driving efficiency

Chances are, your firm has a standard approach to audit planning. And yet, despite proven and time-honored results, auditors are still tripped up on the unexpected: Surprises in the balance sheet, missed variances, insufficient data to ask meaningful questions of management, and more.

How can auditors embrace AI and data analytics to improve audit planning?

It’s time to focus on finding risks rather than executing balance sheet procedures. Join Stuart Cobbe, ACA, and Michael Bottala, CPA for an accountant-driven discussion on how analytics enables an investigative approach to audit planning that cuts to the heart of financial risk:

  • Learn how audit planning is improved by forensic approaches to data

  • Understand different types of analytics, visualizations, and reports

  • See real planning use cases in action

This webinar includes demonstrations and a Q&A session to answer your questions.

You will walk away with the information necessary to decide whether data analytics is right for your audit planning activities.

Transcript

SUMMARY KEYWORDS

audit, transactions, auditors, accounts, firms, approach, data, planning, industry, client, ai, driven, technology, operations, happening, visualizations, bit, business, understand, risk

SPEAKERS

Stuart Cobbe, Michael Bottala

 

Stuart Cobbe 

Thanks, everybody, for joining. We'll be talking today about the role of data driven audit planning within the audit process as a whole, speaking about our technology, speaking about the ways that we're seeing implemented, as well as having a bit of a discussion on the ways that we think it will impact how auditors approach their work, how they gain assurance, and what it means for the future of our profession. So who do you have on the webinar with you today? Michael, I'll let you provide a bit of an introduction to yourself.

 

Michael Bottala 

Yeah, thanks, Stuart. So my name is Michael Botalla. I'm our Director of Sales engineering here in the United States and participate with North America. I've been with MindBridge about a year and a half at this point. And prior to that, I spent four and a half years in public accounting at a regional firm in Los Angeles. You know, obviously, was the boots on the ground before so we're To give you kind of our perspective on how our experience plays into this new technology, you know, we have a lot of interesting capabilities, which I think can really drive board the way we approach an audit.

 

Stuart Cobbe 

Great, thanks. And my name is Stuart Cobb, I'm director of growth for Europe for MindBridge. My background is also an audit. Although I did take a slight divergence into data science and help the mid-tier firm build their analytics team and capability. Today, I'm really focused on helping auditors, accountants and advisors see value from this kind of technology and pushing the boundaries with what's possible using machine learning and AI in this space. So I think before we talk about audit planning before we really talk about what we're using our technology for, in the ways that we're seeing impacting the profession, I think it's important first to set the scene a little bit and talk about our views on the state of audit. I think One of the key things that people at large have really come to realize is that audit is a difficult task, you know, it's quite an undertaking to really understand a business at its most fundamental levels. You know, auditors have to be able to do this very quickly, in increasingly complicated business environment with an increasing number of judgments that need to be made. And they have to formulate a position both on the historic numbers that are contained within the financial statements, as well as often more forward-looking elements such as going concern. And I think, personally speaking, I haven't seen quite a lot of what's coming out of the world of data science. My view is that a lot of the techniques that are in that world will better equip us as auditors to handle the level of complexity and uncertainty that we're facing in our day to day work. And I mean, you know, there's no shortage of headlines about the audit industry. Or the state of financial reporting. I mean, the most recent one and the largest one, really, from a news perspective as, of course, Wirecard. In Germany, huge fraud where a huge amount of money that was reported on the balance sheet just purely didn't exist. And I'm sure most of you are pretty comfortable with these. But nevertheless, I think that there's an opportunity for us to do more as an industry. And this isn't to say that we should be on the backfoot from defending ourselves from allegations that auditors aren't doing enough. But rather there's an opportunity to re-establish ourselves as a source of trust and legitimacy for for companies and for financial statements. And of course, there's constant changing regulation. I mean, you've got ISO 315, was released in is intended to come into action recently. So that's the international standards of auditing for those of us who aren't based in the US, which talks about the increasing Use of these techniques in the planning process. They've also released their intentions to update ISO 500, which is the audit evidence ICER. And, Michael, do you want to talk a little bit about SAS? 142? Over in the US?

 

Michael Bottala 

Yeah, exactly. And that one ties into the audit evidence, you know, utilizing these emerging technologies and the way we can document the analysis that's coming out of these technologies to help support the procedures that were performing the audit process and the conclusions that were coming to. So we're definitely seeing a shift in terms of the, the, the standards in terms of how it's going to be impacting the future of audit. And this is something that can be implemented, you know, early adopters, but it is going to be for December 15th, 2022. So, we are seeing the move towards that it's a gradual movement. But you know, it's something that the industry needs to continue to move forward on.

 

Stuart Cobbe 

Yeah, it's interesting like particularly this automated tools and techniques bullet point, because I find, you know, ultimately audit is one big risk assessment. And I think with some of these techniques, the boundaries between what is a planning technique and what is response to a risk identified at the planning stage, I think in some situations can be a little bit blurred, but our approach is very much to focus on the planning stage to enhance the understanding of the entity at that moment and to use that to drive better outcomes. So, you know, good to see both major regulatory bodies kind of coming to grips and grappling with this. And of course, at the same time, the industry as a whole is coming to accept a lot of these technologies. A study done by the financial reporting Council in the UK, it has found that quite a number of the largest firms are either deploying or piloting usage or researching the use of machine learning. Natural Language Processing, predictive analysis and so on, you know, you can see here and this was a study done in March of 2020. So very, very recent, pretty much all of the firms are exploring the use or actively using one of these technologies in some way, in the UK, and from the FRCS perspective, you know, they're glad to see that there's so much thought being devoted to figuring out how to see value from these kinds of technologies. And, Michael, is it your sense that the same is happening in the US?

 

Michael Bottala 

Yeah, we're definitely seeing a shift towards that, you know, there was obviously the early adopters and firms that, you know, took that innovative approach and have been successful with these technologies. And now we're seeing more firms, you know, try their hand at it and definitely seeing, you know, value in this type of technology in a number of different processes, obviously, the traditional audits but you know, into advisory services into review engagements Hoping to automate a lot of those low cognitive tasks.

 

Stuart Cobbe 

Yeah, it feels a bit like we're kind of crossing that threshold between the early adopters and the early majority for a lot of these technologies, but that's kind of just my, what I'm seeing on the ground. And I mean, that's not to say that people conducting reviews of the auto market aren't aware of what's going on as well. This is a quote from the Brighton review, which is a major review conducted into the state of the audit industry in the UK. And something that sir Donald Bryden noted was the technologies will clearly have an impact on the cost of the audit and the depth of testing that will be possible. So I find both of these quite interesting because they seem this quote really seems to be aware of the fact that data driven techniques involve a more investigatory mindset and more forensic mindset. And will in some instances allow for judgments to be made about whole populations in a much more kind of routine or procedural way. So good to see such important names talking about these topics. And of course, COVID-19 has been a huge impact on the industry as a whole. Three of the themes that we're consistently hearing is that cash is king for many businesses. They're really focused on staying alive during this recession, particularly if they're in the hospitality or travel industry. We've got major publications like the economist talking about how this economic crisis will expose the decade's worth of corporate fraud and I wonder to what extent if it weren't for Coronavirus, we would have seen Wirecard be exposed today. You know, perhaps they would have been able to limp on for a little while longer and There's also an increased opportunity for people to commit fraud. You know, this isn't necessarily corporate fraud, but this is misappropriation of assets or using government funding, perhaps in some other way. So I think with the acceleration of the working from home trend, it's just bringing to light this opportunity for people to do wrong to provide more opportunity to commit fraud on that level. And I think for a lot of these, it's really just an acceleration of trends that we've seen in the industry more broadly. And I found it quite interesting that in the first stages of the Coronavirus crisis, it felt like a lot of firms went into kind of conservation mode. But as firms are charting their strategy for growth post Coronavirus, it feels like artificial intelligence and analytics are a key part of that strategy. And that's certainly been yeah been pretty interesting. I mean, again, you know, Michael, is that something that you In the US?

 

Michael Bottala 

Yeah, definitely. I think when anytime there's a situation like this where you have to reach, you know, these unexpected circumstances come up, a lot of firms have decided, you know, it's a good time to adjust, you know, if we're going to be if there's other unforeseen circumstances in the future, how do we better set ourselves up for those situations and, you know, moving to a more remote atmosphere is obviously one way to do that. And, you know, if there ever is a situation or these things are prolonged, as we're seeing with COVID, you know, what, what everybody thought was going to be a few weeks is turned into months. So, you know, if you adjusted and you are proactive, you've obviously set yourself up better than other firms, and it can become a competitive advantage.

 

Stuart Cobbe 

Yeah, it makes sense. I mean, I always view the Americans as kind of more willing to take risk and to innovate on a lot of these things. And I don't know if that's been the case, but certainly they're better kind of publicizing when they do go out on a limb. So it certainly sticks in the mind a lot more. So, fair enough. I think it's always worth talking about where is AI right now, this is something I like to include in a lot of my presentations. I think it helps people understand where the cutting edge really exists. And what is it that makes this new AI spring so different from the previous times that AI has been a hot topic, I mean, the big change is really the advances in in computational power and the large scale data collection that allows us to train machines to spot patterns. I think for the accounting and auditing industry, it's really the large scale collection of, of data sets in a format that is somewhat structured or that the machine can draw comparisons from across multiple businesses, for instance, is really what's enabling this new era of AI applications in our space. So I think the data side of it is what's really They changing at the moment for auditors. And what about from an academic perspective? What is the performance look like? This is from a great source called the eye index. Again, I include this in all of my presentations, I really recommend for you to check it out. If you're interested in the state of artificial intelligence more broadly, this particular challenge is a visual question answering challenge where the machine must look at pictures that are presented to it and answer a question about what's happening in the scene. So it could be, you know, what is the boy doing and the response that's expected is something around, you know, hitting a baseball with a baseball bat. It's using a very curated data set. So a lot of effort has been put into selecting the right images and giving the right training prompts to the algorithm. And you can see that it's not quite at human baseline performance. There are a few other tasks where the machine has kind of surpassed human baseline performance. This kind of task requires the combination of a lot of high-level reasoning in various different areas. It requires perception. It requires natural language understanding, and so on. So, consistent improvement, but not quite a human baseline. And I also like to draw the comparison to what's going on in the healthcare industry. So this is the story of heart monitoring, really, and I think it's been quite interesting to see how this has developed and we'll draw some parallels to this later on in the presentation. The stage one of heart monitoring was the stethoscope. It's very manual. It was almost an art form to use. It didn't involve no data capture once the process was over and was highly reliant on the doctor being aware and being present during the process. The next phase was really the invention of the electrocardiogram and the standardized data capture that that ran presented, you know, it still requires doctors to have an understanding of what good looks like. And it also requires for doctors to be present during the diagnosis process particularly early in the diagnosis process in order to really screen out those who are healthy or those who need further attention. What we've seen recently is the rise of AI driven diagnostic machines. And this is a study done last year, where it found that an artificial intelligence was able to diagnose problems to the same level of accuracy as human doctors. And what I find particularly interesting about this is it speaks directly to the where the where the medical professionals are spending their time. It enables them to be much more effective with where they're spending their time and really focus on those patients that really need their attention as opposed to the many who possibly to not. So when we're talking about artificial intelligence in the world of audit and accountants, I really like to think of it as three different pillars. The first is AI detection. So similar to the use case that we've just discussed, it's finding that unusual transaction, finding the needle in the haystack, surfacing something that's unusual or risky, or warrants further investigation. There's AI for prediction. So this could be understanding complex relationships, understanding how a business drives its profit line, or how a particular balance is going to perform in the future. And the last element of it is visualization. So taking that information, taking the raw information as well from the business, and making it digestible and understandable for a human. So, I think for me, these are kind of the three key pillars of an artificial intelligence application in our space. And, you know, what's one of the core tenants of MindBridge's AI auditor. We have a risk score and technique of risk scoring algorithm or suite of algorithms that sits under our technology that Michael will show you in a little bit. This really combines domain expertise with statistical methods and with machine learning models. And it's through this combination of different techniques that we can build a robust risk scoring methodology that allows us to surface those transactions or users or classes of transaction that are of interest. This on this idea of an ensemble approach is really taken from the world of data science. You know, it avoids classic problems with relying on a single model to make judgments. And it also improves the accuracy of the techniques, it means that it's more likely to pick up things that are unusual. And so, you know, I just like to cap the section off by saying that There's quite a lot of writing at the moment about how artificial intelligence is coming into the core strategy for many companies, not just within audit. So this is from a BCG paper released in April 2020, the rise of the AI powered company in the post crisis world, and they noted that winners, from this Coronavirus crisis will reinvent themselves by putting software data and AI at the core of their organizations. And we're hoping to show you a little bit about the detail of what that means in our industry in today's presentation.

 

Michael Bottala 

So in this next section, we'll be talking about what planning looks like today. We'll then move into the MindBridge software and talk about the MindBridge approach to altering the current planning, the way you go about planning your audit and then Stuart, we'll get into more of the risk driven audit plan and how that can drive efficiencies in your process. So, let's talk a little bit about where we're at today. So Stuart and I kind of came together and we talked about, you know, what do we do? During the planning stages of the audit, and what we came to understand is it's a very qualitative approach, right, we would review the client history. So we try to understand, you know, what their financial statements look like in the prior year period, we try to have a, you know, an understanding of the industry that they're dealing with, we then would move on to discussions with management. So we'd have conversations with those key employees, those key piece key members of the organization that really drive the operations and understand kind of a holistic view of what's happening within that organization. And then the third piece to it is, you know, there was very basic analytics that were being performed. Typically, there was a year to year variants where we would look for those significant changes and there's very limited number crunching, most of what we would try to discover was based off of more gut feelings and, and kind of a, the the partners experience with this particular client. So the key about current audit planning is if you take the same Approach year to year, you're going to not be able to defer your approach significantly because you're not thinking critically about the data. So, as we move into the historical knowledge and experience with the client, one of the things that is really clear is there's less critical thinking happening because you're relying on your prior experiences. So, you're not taking each year as a unique situation, you're really relying on what you've done in the past and assuming that every year operates the same way. However, you know, there's a lot of situations that come up and this year has been a great example. You know, if you went into the next year audit, you know, planning the same way you have in the past, you're going to, you know, make incorrect assumptions about that particular organization because of the COVID situation this year where things have changed drastically for a lot of different industries.

 

Stuart Cobbe 

And there's also that classic like, you know, moment where the audit partner goes on, don't worry about it, you know, I've we, this is a low risk client, their incomes Really simple, you know, you hear it all the time.

 

Michael Bottala 

Yeah, the other thing, you know, when you're working off of experience and those gut feelings, it's very difficult to document that approach. Right? You know, if the partner tells you No, you as a staff, Stewart, hey, you know, I, I know this from this client in this particular industry, and he goes, go ahead and put that into the planning. You know, you're left a little bit dumbfounded because there's no objective evidence to really support that documentation in your process. Other than referring, you know, the partner said this about this particular client and this engagement. So, I think, you know, moving towards a more data driven approach will lead to more objectivity and the ability to help justify and document your results, which we'll get into in a little bit. But the last key piece to you know, taking that approach with the history of the client is assumptions can be blinders. So if you're assuming that the operations are the same as the past, you could be, you know, incorrectly assuming and kind of turning a blind eye to changes in the organization. we kind of talked about the planning meeting and in management's having discussions with those key stakeholders of the organization, and the way currently, we're approaching the audit with these qualitative approaches, there's no data to ask those more relevant questions about what's happening. So if you as an auditor are going to a CFO, and you're saying, Hey, is there anything I should be aware of in this year, you know, the CFO is can be very clear, or the comptroller or accounting manager, those key personnel are going to, you know, give you very basic answers, such as very little happened in the year, you know, how is business this year? Oh, business was good. Can you tell me a little bit about what happened? Well, sales are up, you know, things are operating, we're moving along very well. And it's really a client driven conversation versus the auditor driving the conversation because you're trying to poke and prod and trying to figure out what's happening but you're not getting the responses that you're truly looking for.

 

Stuart Cobbe 

Well, it's just In the fundamental mismatch and the amount of information that both of them have, you know, like the client just fundamentally knows much more. And the auditor is starting from a, from a lower position for, for what's happened.

 

Michael Bottala 

Yeah, that's a great point. And then the last key piece to kind of the current planning stages is about how we approach analytics in the in the planning stages using very limited data. So, we can see here, this is a significant change in a couple of accounts. For this particular client, we can see that the accounts receivable how to 27% change, and the comment is income is up for 2017. Now that doesn't tell us very much about what's particularly happening in accounts receivable. So how do we design our audit plan, if we have very basic knowledge of a particular account, we would need to understand more about what's happening over the course of the year and whether this change is part of normal course of operations or if there's significant fluctuations that we haven't accounted for another if we move down one line there To the cache, which we have an example of, we can see a 9% change in the cache. But once again, that's not a very significant change, you know, it's under 10%. So this is something that we could overlook. And say it's not significant to the financials with our immaterial change comment, because everything looks fine. But if we move into the next slide, we'll see, you know, if we incorporate more data into this approach, where now we have four years of information, we can see how the cash account is acting over this period of time. So we can see it's been consistently falling from 2014 to through 2017. In a very similar pattern to taking that approach, we now we can start digging into why this is consistently falling, which, which parts of the operations are really impacting cash. What we can also see is cash gets extremely low in the month of June, and maybe there's a cause for concern in terms of cash flow available to pay our vendors to keep up operations of flow. So, it's giving us more perspective into going concern as well, you know if that if this continues to get lower in the future, is that going to, you know, impact our operations negatively. Let's go ahead and hop into the MindBridge platform and talk about how if we take that same approach to our audit process, using more data, and analytics, we can definitely improve and gain more insights into the operations. So first of all, let's describe a little bit about what MindBridge does MindBridge, in just a general ledger AR or AP sub ledger, we analyze 100% of the transactions using our control points. So these control points are a mixture of business rules, statistical models and machine learning, which Stuart mentioned earlier, and they all have different weightings assigned to them. As the control point is being triggered. We're building out a risk score for all these different underlying transactions which are then classified in To three key areas, high risk, which is greater than 50% risk medium, which is between 30 and 50%. And low, which is less than 30% risk. If we then dive into a particular transaction, we're able to see the line items that make up that transaction as well as the control points that are being triggered down below. So to talk a little bit about these control points, and how we differ from the traditional tools that are out there is really around our approach with the traditional tools. They're, they're based off of business rules and statistical models. And and in order to develop those rules you need to know what you're looking for. Whereas with Muybridge, we're really able to uncover those unknowns in the data set because we're looking for those unusual patterns in the data. And it's not necessarily geared towards a specific test. So with that approach, we feel that we're able to really identify those most unusual and most uncommon transactions that could have a material impact on your family. Annual statements. And the way we do that is with control points such as outlier anomaly. So, the way outlier anomaly works is we look for the relationships between the accounts, and we develop a neighborhood and what's normal in terms of the announcer accorded to that particular transaction. So, for example, if you recorded a debit to your repairs and maintenance and a credit to your accounts payable for that particular relationship, if the transactions were usually around 200 and $500, we're starting to develop a normal kind of neighborhood in terms of where the amounts lie. Similar to outlier anomaly in terms of looking at the relationships, we we've also incorporated rare flows. So rare flows, looks at, once again, the relationship but how frequent that relationships occurring within the data set. So we're able to understand, you know, Is this normal and common transaction or is this something that's a one off and if this one off situation can improve Back to financial statements. So these two are really able to uncover those unknowns within the data set. The third key component of that is building a domain expertise with what we call expert score. So experts score was developed by surveying auditors and accountants and having them risk or different types of relationships. So because of this, we're able to flag those most unusual relationships that don't fall into typical accounting, you know, in the US what we call gap in trying to understand, you know, what's a normal relationship. So taking this approach, once again, all these different tests are layered on top of each other in order to develop that risk score. So if it's a duplicate transaction, it's a high monetary value, and it's a rare flow, it's going to be a riskier transaction than if it just one of these control points was being triggered. And that's really the core of the tool is to help surface these items can give you visibility to enhance your audit approach. So Now that we have a basic understanding of how this system operates, what we'll do is we'll kind of comb through the functionality and how it can be used in different areas of the audit. So the first piece is the risk overview, the risk overview, once first of all provides a nice high level overview of all the data coming through the general ledger for this particular client, it breaks it down over time by account as well as by different parameters such as who's entering those transactions, what state those transactions are related to our cost center, those transactions are related to. From here, we can use our qualitative understanding to start digging into those key accounts that we know about for this particular client. So the first thing that we could do is we can dig into an account such as our cash, and we can say, Okay, what are higher risk transactions coming through cash are these things that we're aware of or these things that could impact The financial statements and we need to dig into further. Once again, getting that visibility beforehand. That way we can account for it in our audit plan and gain more efficiencies on the back end, which Stewart will dig into later. But what's really nice about this is now this whole page is updated related to this particular account. So, we can see that, you know, the stratification of our entries over time, we can see there's a higher volume in the beginning of the year versus the end of the year. So maybe there's a little bit of seasonality to the product that we're selling. In this particular client, what we're also able to gain visibility into is the control points being triggered. So these control points can definitely impact the different assertions that we're looking at. So, for example, if we have a number of transactions that are being recorded and triggering our end of year control point, maybe there's an issue with the cutoff in the organization, or maybe there's a number of sequence gaps coming through our cash account. Maybe there's an issue with the completeness of the data or you know, tying into those unknowns, maybe there's a lot of rare flows and unusual amounts being recorded that that factor into the existence of these transactions and if they're legitimate or not. So taking this approach, we're able to gain a lot more understanding of the type of transactions that are coming through. What we can also do is we can filter off of the type of transaction. So if we select the type of transaction, we can then dig into our cost of goods sold, for example. And what this allows us to do is really understand the controls in the organization. So we can see, you know, this risk by account has been updated based off of this type of entry. So if we're seeing, you know, inventory and our cost of materials, this is in line with our expectations, but if our cost of goods sold was hitting a sales account hitting expense accounts or other balance sheet accounts that we weren't aware of. Now we have visibility into how these entries are being recorded and if they have proper controls within the organization and now we can justify, you know, if everything is being recorded the way we expected it to. Maybe we want to reduce the controls in this particular year because we're confident that the transactions are hitting the correct accounts. As we scroll over to the right, we can also see what users are entering those Cost of Goods Sold transactions. Maybe there's particular users that shouldn't be recording transactions in this area. Or maybe there's a lot of management override in this particular area. This is some things that we can gain visibility into. And now we can dig into we can see there's a bit an outlier here with 15 user 1578. Now we're able to drill down into that user and get a better understanding if the controls are operating as they should. So, as you can see, we can get a nice we can we can take a more objective approach and really quantify the type of transactions and really understand you know, the the way These entries can impact the financials. From here, we can then bolster our preliminary analytics. And the first way to do that is taking a similar approach to what we've done in the past in terms of looking at our significant variances, so it's still a crucial step because it's a nice starting point or any of these accounts that maybe had significant changes from the prior period. So right now we're looking at a horizontal analysis of our balance sheet, we can also look at it based off the income statement view. And then we can look at a vertical analysis, which is common size financials of these different accounts and how that compares to the prior period to give you more perspective on how these particular accounts compared to the prior period. So, once we have an understanding of those accounts, we can then drill down into our trending section and look into those particular accounts. So maybe we want to break down our current assets and see what that looks like over the course of the year. What we can also do which is unique to Muybridge is look at the activity. So looking at the ending balance is one thing, but looking at the activity really gives you that perspective on how the operations are being affected by these by the beat by the underlying entries, and whether the operations are consistent with prior periods. So if we get a little bit more specific, and we drill into our accounts receivable, for example, you know, we can look at we can do, we can click off a few years to to get a nice, cleaner picture and we can see, you know, in 2000, in the prior period, we had a big dip in the month of October, but it's fairly consistent in the current period. So what happened in October that didn't happen in October of this current period, and now we can document and change our approach accordingly. Lastly, you know, what we can do is we can dig into the, the different ratios that are key to the organization and highlight those time periods. The ratios do not meet our expectations. So within Muybridge, we have an algorithm called Serima, which takes an integrated moving average of the data to develop an expected range. So, from this expected range, we're able to compare that to the actual results. And if there's anything that falls outside of our expectations, we'll highlight it with a red outlier circle here. So, what that allows us to do is really target now this is days expenses in AP. So now we can target these particular accounts. And we can drill down into how these could affect this this particular month could affect the financials because it wasn't within the normal operations and within our expectations. So now we can adjust our audit plan to to really consider the timing of our day’s expenses in AP for the month of July. What's also unique to Muybridge is the ability to build out all these ratios beforehand. So we mentioned how AI can help automate a lot of those low cognitive tasks. One of the ways that we can do that is by producing these visualizations automatically for you. So, by having these here and ready to go, it's just a matter of interpreting the visualizations now, and annotating or making your comments on these visualizations. And then building out a nice work paper to support the process that you're taking within Muybridge. But once again, all this is displayed for you. So, you don't need to worry about building out these trends building out these ratios, it's just a matter of looking into how it can these, these can affect the current year audit that we're doing. So, from here, we can then go into the data table where we can drill down into those transactions and performer substantive procedures. But the focus for today is on audit planning. So, we'll dive back into our slide deck. This kind of summarizes everything that we looked at within the Muybridge platform. Right. The key is we're driving better internal and external communication through this quantitative approach. The way we're doing this is by flagging those unusual transactions and understanding how these unusual transactions can impact the financial statements. We're getting a deeper understanding the classes of transactions and how those impact our controls and the operations. And then lastly, our enhanced analytics. So these trends and ratios once we start incorporating more data, we're getting a better picture of what's actually happening within the organization. And now we're taking more specific questions to the client. So you can see how, you know the approach of the past can be enhanced through a tool like this to really develop that feedback loop in a quantitative analysis to be paired with the qualitative analysis that you've done in the past

 

Stuart Cobbe 

some of our thoughts on why this matters, as well as some examples inspired by real life situations that that kind of bring some of this into a little bit more context and why we think it's important from the perspective of driving an efficient audit. You know, the first is that I believe strongly in the value of an early warning system. And this is a great example taken from Dan Heath book upstream that talks all about how to solve problems before they happen. This is an example taken from Northwell Health, which is a health provider in the New York State area, and their emfs. Their Emergency Services Office used data used a lot of history of 911 calls to predict when and why they were happening in order to position their ambulances in various spots around the city so as to react quicker. And they ended up shaving off about a minute and a half of their response time. And you can see here that they perform significantly better than the US national average. And the perspective of Dan was that when you can foresee a problem, you have more maneuvering room to fix it. And I can think of a number of examples of when I was in audit were issues that would crop up late in the Process had this cascading effect whereby all of a sudden you had to pull in a more expensive manager, or you had to unpick a lot of the work that was done previously. I think the advantage of starting an audit with a clear data driven understanding of where problems might particularly lie within a business means it's far more likely that you'll have these kind of toppling dominoes at the end of the audit process where everybody's pulling their hair out. To give you an example, inspired by real life, this is an example of those kinds of visualizations that Michael is showing you in the tool. This is filtered for the manual adjustment, and on the left, we've got the risk burst. So, each segment of this donut represents a part of the chart of accounts that is being affected by this transaction type. The size of this segment represents the value of transactions being posted and the color represents the average risk and you can see there's a lot of Highly material movements going on in the profit and loss on foreign exchange account and these are being posted with manual journals. So, you know instantly as an auditor at the planning stage of the audit, I've got what I would consider to be quite a big red flag that needs to be waived. And I should also be a bit concerned about the control environment that's allowing somebody to post these transactions. And be I should be scheduling time to have somebody who is relatively senior and understands the complexity of foreign exchange, to be doing the work on that section of the audit because I know it might be a problem. And all of that is just reinforced by the visualization from the trending dashboard. When you look at the activity that's touching that profit or loss and foreign exchange, you see these huge swings up and down, you know, highly material movements. All of it seems to be reversed and then posted in the last period of the year. Seems pretty strange, pretty suspicious. Or possibly pretty unusual. And so, you know, the second element apart from this early warning system is really taking judgments that are based on evidence rather than on gut feel and putting engagement teams in a position where they can put that evidence down on the audit file. To use another quote from the Bridon review, regardless of the Genesis sample sizes today, oh, more to audit partners judgments than to statistical analysis. And again, this is something that I would largely agree with, from what I've seen in the audit industry. And I think what we're starting to see now is a shift towards more evidence-based justifications for sample sizes. We're seeing regulators taking a pretty dim view on kept sample sizes, for instance. So all of a sudden, the need for auditors to justify why they're doing a particular amount of work is becoming much more necessary. And so, it is An example of that closer to real life than what Michael had shown you before. This is a purchase invoices being posted, you know, it's touching largely trade accounts payable. It's going between the nominal accounts that we expect for this particular business, the vast majority of the value sits within low risk, the seasonality of the postings matches the seasonality of the business as a whole. And it's all being posted by members of the team that we expect to be posting these transactions with a low average risk. So we can see how we're building up an evidence based view of what's going on within this particular class of, of transaction. And so, you know, using both of these, all of a sudden, the planning meeting takes a slightly different slant to it. We've got client driven planning meeting, but really, we meant auditor driven planning meeting where the auditor is armed with the information, they need in order to ask the really incisive questions at the planning stage. You know, why are there so many manual adjustments made to the cash account? And this is happening much earlier in the process when the CFO and management are much more receptive to these difficult questions being asked. It also means that the preparation for that planning meeting is a team effort. It's not just the partner rocking up with their clipboard and asking questions. we're enabling other members of the team to feed into the questions that are being generated for that planning meeting. Which I think only means that the audit team as a whole is better prepared. And all of this is really being reflected in our clients. So there's this great quote from Baldwin's MindBridge's driving better conversations with our clients, which drives value the responses being you're asking better questions, which is exactly what we're trying to achieve. And that's not to say that we'll always get to the bottom of the problem in the first instance, but It certainly means that you've got a head start when you've got this kind of race against time that many audits represent. So what does it take to implement?

 

Michael Bottala 

So, what it really comes down to is as a cultural change, right, it's about altering your approach in order to take more collaborative and interdisciplinary approach to your audit plan. As Stewart just mentioned, with incorporating multiple different team members, and having them make more of an impact, in terms of the way you're approaching the audit. You know, instead of taking all these gut feelings in the experience base that we talked about earlier, we're using this data to take a more quantitative approach in order to really drive decision making. And what that allows you to do is it allows you to incorporate the younger individuals within the firm because they have the ability to interpret this data and bring value to the table and it's not all driven from the top down. You know, here's, here's the thing. Different things that have just been kind of quoted by the IAASB, you know, in regard to these different changes, so the rate of change and increasing use of big data capabilities and knowledge, you know, these are all common themes within our, our industry. And it's, you know, the time is now it's this isn't a new technology anymore. It's really something that's very prevalent and firms are taking advantage of it, such as Baldwin, CPA, missing any of these key elements in this change scorecard. It can lead to confusion; it can lead to anxiety or resistance from your team. And definitely a little bit of frustration. You know, anytime you're implementing a new process, there's going to be frustration, but the goal is to understand why you're doing it and why you're trying to get to where you're going with technology such as this. So keep in mind, you know, we want to make sure that we have a clear and definitive plan with our vision skills, incentive resources, and an action all aligned in order to really implement that change.

 

Stuart Cobbe 

To be honest, what I've what I've really seen the most is, is where the lack of resources. I mean, I think a lot of these things require an investment. And they really require people to have the dedicated time and headspace to think about what this means and whether it's coming up with the planning template or looking at how the communications need to change to the client prior to the planning meeting. There are elements that firms will need to think about and obviously we do everything we can to help, but nevertheless, the firm needs to be prepared. Thank you for coming, everybody. Hopefully you found it interesting. And feel free to get in touch with any of us here MindBridge if you've got more questions. Have a good day.

SUMMARY KEYWORDS

audit, transactions, auditors, accounts, firms, approach, data, planning, industry, client, ai, driven, technology, operations, happening, visualizations, bit, business, understand, risk

SPEAKERS

Stuart Cobbe, Michael Bottala

Stuart Cobbe 

Thanks, everybody, for joining. We'll be talking today about the role of data driven audit planning within the audit process as a whole, speaking about our technology, speaking about the ways that we're seeing implemented, as well as having a bit of a discussion on the ways that we think it will impact how auditors approach their work, how they gain assurance, and what it means for the future of our profession. So who do you have on the webinar with you today? Michael, I'll let you provide a bit of an introduction to yourself.

 

Michael Bottala 

Yeah, thanks, Stuart. So my name is Michael Botalla. I'm our Director of Sales engineering here in the United States and participate with North America. I've been with MindBridge about a year and a half at this point. And prior to that, I spent four and a half years in public accounting at a regional firm in Los Angeles. You know, obviously, was the boots on the ground before so we're To give you kind of our perspective on how our experience plays into this new technology, you know, we have a lot of interesting capabilities, which I think can really drive board the way we approach an audit.

 

Stuart Cobbe 

Great, thanks. And my name is Stuart Cobb, I'm director of growth for Europe for MindBridge. My background is also an audit. Although I did take a slight divergence into data science and help the mid-tier firm build their analytics team and capability. Today, I'm really focused on helping auditors, accountants and advisors see value from this kind of technology and pushing the boundaries with what's possible using machine learning and AI in this space. So I think before we talk about audit planning before we really talk about what we're using our technology for, in the ways that we're seeing impacting the profession, I think it's important first to set the scene a little bit and talk about our views on the state of audit. I think One of the key things that people at large have really come to realize is that audit is a difficult task, you know, it's quite an undertaking to really understand a business at its most fundamental levels. You know, auditors have to be able to do this very quickly, in increasingly complicated business environment with an increasing number of judgments that need to be made. And they have to formulate a position both on the historic numbers that are contained within the financial statements, as well as often more forward-looking elements such as going concern. And I think, personally speaking, I haven't seen quite a lot of what's coming out of the world of data science. My view is that a lot of the techniques that are in that world will better equip us as auditors to handle the level of complexity and uncertainty that we're facing in our day to day work. And I mean, you know, there's no shortage of headlines about the audit industry. Or the state of financial reporting. I mean, the most recent one and the largest one, really, from a news perspective as, of course, Wirecard. In Germany, huge fraud where a huge amount of money that was reported on the balance sheet just purely didn't exist. And I'm sure most of you are pretty comfortable with these. But nevertheless, I think that there's an opportunity for us to do more as an industry. And this isn't to say that we should be on the backfoot from defending ourselves from allegations that auditors aren't doing enough. But rather there's an opportunity to re-establish ourselves as a source of trust and legitimacy for for companies and for financial statements. And of course, there's constant changing regulation. I mean, you've got ISO 315, was released in is intended to come into action recently. So that's the international standards of auditing for those of us who aren't based in the US, which talks about the increasing Use of these techniques in the planning process. They've also released their intentions to update ISO 500, which is the audit evidence ICER. And, Michael, do you want to talk a little bit about SAS? 142? Over in the US?

 

Michael Bottala 

Yeah, exactly. And that one ties into the audit evidence, you know, utilizing these emerging technologies and the way we can document the analysis that's coming out of these technologies to help support the procedures that were performing the audit process and the conclusions that were coming to. So we're definitely seeing a shift in terms of the, the, the standards in terms of how it's going to be impacting the future of audit. And this is something that can be implemented, you know, early adopters, but it is going to be for December 15th, 2022. So, we are seeing the move towards that it's a gradual movement. But you know, it's something that the industry needs to continue to move forward on.

 

Stuart Cobbe 

Yeah, it's interesting like particularly this automated tools and techniques bullet point, because I find, you know, ultimately audit is one big risk assessment. And I think with some of these techniques, the boundaries between what is a planning technique and what is response to a risk identified at the planning stage, I think in some situations can be a little bit blurred, but our approach is very much to focus on the planning stage to enhance the understanding of the entity at that moment and to use that to drive better outcomes. So, you know, good to see both major regulatory bodies kind of coming to grips and grappling with this. And of course, at the same time, the industry as a whole is coming to accept a lot of these technologies. A study done by the financial reporting Council in the UK, it has found that quite a number of the largest firms are either deploying or piloting usage or researching the use of machine learning. Natural Language Processing, predictive analysis and so on, you know, you can see here and this was a study done in March of 2020. So very, very recent, pretty much all of the firms are exploring the use or actively using one of these technologies in some way, in the UK, and from the FRCS perspective, you know, they're glad to see that there's so much thought being devoted to figuring out how to see value from these kinds of technologies. And, Michael, is it your sense that the same is happening in the US?

 

Michael Bottala 

Yeah, we're definitely seeing a shift towards that, you know, there was obviously the early adopters and firms that, you know, took that innovative approach and have been successful with these technologies. And now we're seeing more firms, you know, try their hand at it and definitely seeing, you know, value in this type of technology in a number of different processes, obviously, the traditional audits but you know, into advisory services into review engagements Hoping to automate a lot of those low cognitive tasks.

 

Stuart Cobbe 

Yeah, it feels a bit like we're kind of crossing that threshold between the early adopters and the early majority for a lot of these technologies, but that's kind of just my, what I'm seeing on the ground. And I mean, that's not to say that people conducting reviews of the auto market aren't aware of what's going on as well. This is a quote from the Brighton review, which is a major review conducted into the state of the audit industry in the UK. And something that sir Donald Bryden noted was the technologies will clearly have an impact on the cost of the audit and the depth of testing that will be possible. So I find both of these quite interesting because they seem this quote really seems to be aware of the fact that data driven techniques involve a more investigatory mindset and more forensic mindset. And will in some instances allow for judgments to be made about whole populations in a much more kind of routine or procedural way. So good to see such important names talking about these topics. And of course, COVID-19 has been a huge impact on the industry as a whole. Three of the themes that we're consistently hearing is that cash is king for many businesses. They're really focused on staying alive during this recession, particularly if they're in the hospitality or travel industry. We've got major publications like the economist talking about how this economic crisis will expose the decade's worth of corporate fraud and I wonder to what extent if it weren't for Coronavirus, we would have seen Wirecard be exposed today. You know, perhaps they would have been able to limp on for a little while longer and There's also an increased opportunity for people to commit fraud. You know, this isn't necessarily corporate fraud, but this is misappropriation of assets or using government funding, perhaps in some other way. So I think with the acceleration of the working from home trend, it's just bringing to light this opportunity for people to do wrong to provide more opportunity to commit fraud on that level. And I think for a lot of these, it's really just an acceleration of trends that we've seen in the industry more broadly. And I found it quite interesting that in the first stages of the Coronavirus crisis, it felt like a lot of firms went into kind of conservation mode. But as firms are charting their strategy for growth post Coronavirus, it feels like artificial intelligence and analytics are a key part of that strategy. And that's certainly been yeah been pretty interesting. I mean, again, you know, Michael, is that something that you In the US?

 

Michael Bottala 

Yeah, definitely. I think when anytime there's a situation like this where you have to reach, you know, these unexpected circumstances come up, a lot of firms have decided, you know, it's a good time to adjust, you know, if we're going to be if there's other unforeseen circumstances in the future, how do we better set ourselves up for those situations and, you know, moving to a more remote atmosphere is obviously one way to do that. And, you know, if there ever is a situation or these things are prolonged, as we're seeing with COVID, you know, what, what everybody thought was going to be a few weeks is turned into months. So, you know, if you adjusted and you are proactive, you've obviously set yourself up better than other firms, and it can become a competitive advantage.

 

Stuart Cobbe 

Yeah, it makes sense. I mean, I always view the Americans as kind of more willing to take risk and to innovate on a lot of these things. And I don't know if that's been the case, but certainly they're better kind of publicizing when they do go out on a limb. So it certainly sticks in the mind a lot more. So, fair enough. I think it's always worth talking about where is AI right now, this is something I like to include in a lot of my presentations. I think it helps people understand where the cutting edge really exists. And what is it that makes this new AI spring so different from the previous times that AI has been a hot topic, I mean, the big change is really the advances in in computational power and the large scale data collection that allows us to train machines to spot patterns. I think for the accounting and auditing industry, it's really the large scale collection of, of data sets in a format that is somewhat structured or that the machine can draw comparisons from across multiple businesses, for instance, is really what's enabling this new era of AI applications in our space. So I think the data side of it is what's really They changing at the moment for auditors. And what about from an academic perspective? What is the performance look like? This is from a great source called the eye index. Again, I include this in all of my presentations, I really recommend for you to check it out. If you're interested in the state of artificial intelligence more broadly, this particular challenge is a visual question answering challenge where the machine must look at pictures that are presented to it and answer a question about what's happening in the scene. So it could be, you know, what is the boy doing and the response that's expected is something around, you know, hitting a baseball with a baseball bat. It's using a very curated data set. So a lot of effort has been put into selecting the right images and giving the right training prompts to the algorithm. And you can see that it's not quite at human baseline performance. There are a few other tasks where the machine has kind of surpassed human baseline performance. This kind of task requires the combination of a lot of high-level reasoning in various different areas. It requires perception. It requires natural language understanding, and so on. So, consistent improvement, but not quite a human baseline. And I also like to draw the comparison to what's going on in the healthcare industry. So this is the story of heart monitoring, really, and I think it's been quite interesting to see how this has developed and we'll draw some parallels to this later on in the presentation. The stage one of heart monitoring was the stethoscope. It's very manual. It was almost an art form to use. It didn't involve no data capture once the process was over and was highly reliant on the doctor being aware and being present during the process. The next phase was really the invention of the electrocardiogram and the standardized data capture that that ran presented, you know, it still requires doctors to have an understanding of what good looks like. And it also requires for doctors to be present during the diagnosis process particularly early in the diagnosis process in order to really screen out those who are healthy or those who need further attention. What we've seen recently is the rise of AI driven diagnostic machines. And this is a study done last year, where it found that an artificial intelligence was able to diagnose problems to the same level of accuracy as human doctors. And what I find particularly interesting about this is it speaks directly to the where the where the medical professionals are spending their time. It enables them to be much more effective with where they're spending their time and really focus on those patients that really need their attention as opposed to the many who possibly to not. So when we're talking about artificial intelligence in the world of audit and accountants, I really like to think of it as three different pillars. The first is AI detection. So similar to the use case that we've just discussed, it's finding that unusual transaction, finding the needle in the haystack, surfacing something that's unusual or risky, or warrants further investigation. There's AI for prediction. So this could be understanding complex relationships, understanding how a business drives its profit line, or how a particular balance is going to perform in the future. And the last element of it is visualization. So taking that information, taking the raw information as well from the business, and making it digestible and understandable for a human. So, I think for me, these are kind of the three key pillars of an artificial intelligence application in our space. And, you know, what's one of the core tenants of MindBridge's AI auditor. We have a risk score and technique of risk scoring algorithm or suite of algorithms that sits under our technology that Michael will show you in a little bit. This really combines domain expertise with statistical methods and with machine learning models. And it's through this combination of different techniques that we can build a robust risk scoring methodology that allows us to surface those transactions or users or classes of transaction that are of interest. This on this idea of an ensemble approach is really taken from the world of data science. You know, it avoids classic problems with relying on a single model to make judgments. And it also improves the accuracy of the techniques, it means that it's more likely to pick up things that are unusual. And so, you know, I just like to cap the section off by saying that There's quite a lot of writing at the moment about how artificial intelligence is coming into the core strategy for many companies, not just within audit. So this is from a BCG paper released in April 2020, the rise of the AI powered company in the post crisis world, and they noted that winners, from this Coronavirus crisis will reinvent themselves by putting software data and AI at the core of their organizations. And we're hoping to show you a little bit about the detail of what that means in our industry in today's presentation.

 

Michael Bottala 

So in this next section, we'll be talking about what planning looks like today. We'll then move into the MindBridge software and talk about the MindBridge approach to altering the current planning, the way you go about planning your audit and then Stuart, we'll get into more of the risk driven audit plan and how that can drive efficiencies in your process. So, let's talk a little bit about where we're at today. So Stuart and I kind of came together and we talked about, you know, what do we do? During the planning stages of the audit, and what we came to understand is it's a very qualitative approach, right, we would review the client history. So we try to understand, you know, what their financial statements look like in the prior year period, we try to have a, you know, an understanding of the industry that they're dealing with, we then would move on to discussions with management. So we'd have conversations with those key employees, those key piece key members of the organization that really drive the operations and understand kind of a holistic view of what's happening within that organization. And then the third piece to it is, you know, there was very basic analytics that were being performed. Typically, there was a year to year variants where we would look for those significant changes and there's very limited number crunching, most of what we would try to discover was based off of more gut feelings and, and kind of a, the the partners experience with this particular client. So the key about current audit planning is if you take the same Approach year to year, you're going to not be able to defer your approach significantly because you're not thinking critically about the data. So, as we move into the historical knowledge and experience with the client, one of the things that is really clear is there's less critical thinking happening because you're relying on your prior experiences. So, you're not taking each year as a unique situation, you're really relying on what you've done in the past and assuming that every year operates the same way. However, you know, there's a lot of situations that come up and this year has been a great example. You know, if you went into the next year audit, you know, planning the same way you have in the past, you're going to, you know, make incorrect assumptions about that particular organization because of the COVID situation this year where things have changed drastically for a lot of different industries.

 

Stuart Cobbe 

And there's also that classic like, you know, moment where the audit partner goes on, don't worry about it, you know, I've we, this is a low risk client, their incomes Really simple, you know, you hear it all the time.

 

Michael Bottala 

Yeah, the other thing, you know, when you're working off of experience and those gut feelings, it's very difficult to document that approach. Right? You know, if the partner tells you No, you as a staff, Stewart, hey, you know, I, I know this from this client in this particular industry, and he goes, go ahead and put that into the planning. You know, you're left a little bit dumbfounded because there's no objective evidence to really support that documentation in your process. Other than referring, you know, the partner said this about this particular client and this engagement. So, I think, you know, moving towards a more data driven approach will lead to more objectivity and the ability to help justify and document your results, which we'll get into in a little bit. But the last key piece to you know, taking that approach with the history of the client is assumptions can be blinders. So if you're assuming that the operations are the same as the past, you could be, you know, incorrectly assuming and kind of turning a blind eye to changes in the organization. we kind of talked about the planning meeting and in management's having discussions with those key stakeholders of the organization, and the way currently, we're approaching the audit with these qualitative approaches, there's no data to ask those more relevant questions about what's happening. So if you as an auditor are going to a CFO, and you're saying, Hey, is there anything I should be aware of in this year, you know, the CFO is can be very clear, or the comptroller or accounting manager, those key personnel are going to, you know, give you very basic answers, such as very little happened in the year, you know, how is business this year? Oh, business was good. Can you tell me a little bit about what happened? Well, sales are up, you know, things are operating, we're moving along very well. And it's really a client driven conversation versus the auditor driving the conversation because you're trying to poke and prod and trying to figure out what's happening but you're not getting the responses that you're truly looking for.

 

Stuart Cobbe 

Well, it's just In the fundamental mismatch and the amount of information that both of them have, you know, like the client just fundamentally knows much more. And the auditor is starting from a, from a lower position for, for what's happened.

 

Michael Bottala 

Yeah, that's a great point. And then the last key piece to kind of the current planning stages is about how we approach analytics in the in the planning stages using very limited data. So, we can see here, this is a significant change in a couple of accounts. For this particular client, we can see that the accounts receivable how to 27% change, and the comment is income is up for 2017. Now that doesn't tell us very much about what's particularly happening in accounts receivable. So how do we design our audit plan, if we have very basic knowledge of a particular account, we would need to understand more about what's happening over the course of the year and whether this change is part of normal course of operations or if there's significant fluctuations that we haven't accounted for another if we move down one line there To the cache, which we have an example of, we can see a 9% change in the cache. But once again, that's not a very significant change, you know, it's under 10%. So this is something that we could overlook. And say it's not significant to the financials with our immaterial change comment, because everything looks fine. But if we move into the next slide, we'll see, you know, if we incorporate more data into this approach, where now we have four years of information, we can see how the cash account is acting over this period of time. So we can see it's been consistently falling from 2014 to through 2017. In a very similar pattern to taking that approach, we now we can start digging into why this is consistently falling, which, which parts of the operations are really impacting cash. What we can also see is cash gets extremely low in the month of June, and maybe there's a cause for concern in terms of cash flow available to pay our vendors to keep up operations of flow. So, it's giving us more perspective into going concern as well, you know if that if this continues to get lower in the future, is that going to, you know, impact our operations negatively. Let's go ahead and hop into the MindBridge platform and talk about how if we take that same approach to our audit process, using more data, and analytics, we can definitely improve and gain more insights into the operations. So first of all, let's describe a little bit about what MindBridge does MindBridge, in just a general ledger AR or AP sub ledger, we analyze 100% of the transactions using our control points. So these control points are a mixture of business rules, statistical models and machine learning, which Stuart mentioned earlier, and they all have different weightings assigned to them. As the control point is being triggered. We're building out a risk score for all these different underlying transactions which are then classified in To three key areas, high risk, which is greater than 50% risk medium, which is between 30 and 50%. And low, which is less than 30% risk. If we then dive into a particular transaction, we're able to see the line items that make up that transaction as well as the control points that are being triggered down below. So to talk a little bit about these control points, and how we differ from the traditional tools that are out there is really around our approach with the traditional tools. They're, they're based off of business rules and statistical models. And and in order to develop those rules you need to know what you're looking for. Whereas with Muybridge, we're really able to uncover those unknowns in the data set because we're looking for those unusual patterns in the data. And it's not necessarily geared towards a specific test. So with that approach, we feel that we're able to really identify those most unusual and most uncommon transactions that could have a material impact on your family. Annual statements. And the way we do that is with control points such as outlier anomaly. So, the way outlier anomaly works is we look for the relationships between the accounts, and we develop a neighborhood and what's normal in terms of the announcer accorded to that particular transaction. So, for example, if you recorded a debit to your repairs and maintenance and a credit to your accounts payable for that particular relationship, if the transactions were usually around 200 and $500, we're starting to develop a normal kind of neighborhood in terms of where the amounts lie. Similar to outlier anomaly in terms of looking at the relationships, we we've also incorporated rare flows. So rare flows, looks at, once again, the relationship but how frequent that relationships occurring within the data set. So we're able to understand, you know, Is this normal and common transaction or is this something that's a one off and if this one off situation can improve Back to financial statements. So these two are really able to uncover those unknowns within the data set. The third key component of that is building a domain expertise with what we call expert score. So experts score was developed by surveying auditors and accountants and having them risk or different types of relationships. So because of this, we're able to flag those most unusual relationships that don't fall into typical accounting, you know, in the US what we call gap in trying to understand, you know, what's a normal relationship. So taking this approach, once again, all these different tests are layered on top of each other in order to develop that risk score. So if it's a duplicate transaction, it's a high monetary value, and it's a rare flow, it's going to be a riskier transaction than if it just one of these control points was being triggered. And that's really the core of the tool is to help surface these items can give you visibility to enhance your audit approach. So Now that we have a basic understanding of how this system operates, what we'll do is we'll kind of comb through the functionality and how it can be used in different areas of the audit. So the first piece is the risk overview, the risk overview, once first of all provides a nice high level overview of all the data coming through the general ledger for this particular client, it breaks it down over time by account as well as by different parameters such as who's entering those transactions, what state those transactions are related to our cost center, those transactions are related to. From here, we can use our qualitative understanding to start digging into those key accounts that we know about for this particular client. So the first thing that we could do is we can dig into an account such as our cash, and we can say, Okay, what are higher risk transactions coming through cash are these things that we're aware of or these things that could impact The financial statements and we need to dig into further. Once again, getting that visibility beforehand. That way we can account for it in our audit plan and gain more efficiencies on the back end, which Stewart will dig into later. But what's really nice about this is now this whole page is updated related to this particular account. So, we can see that, you know, the stratification of our entries over time, we can see there's a higher volume in the beginning of the year versus the end of the year. So maybe there's a little bit of seasonality to the product that we're selling. In this particular client, what we're also able to gain visibility into is the control points being triggered. So these control points can definitely impact the different assertions that we're looking at. So, for example, if we have a number of transactions that are being recorded and triggering our end of year control point, maybe there's an issue with the cutoff in the organization, or maybe there's a number of sequence gaps coming through our cash account. Maybe there's an issue with the completeness of the data or you know, tying into those unknowns, maybe there's a lot of rare flows and unusual amounts being recorded that that factor into the existence of these transactions and if they're legitimate or not. So taking this approach, we're able to gain a lot more understanding of the type of transactions that are coming through. What we can also do is we can filter off of the type of transaction. So if we select the type of transaction, we can then dig into our cost of goods sold, for example. And what this allows us to do is really understand the controls in the organization. So we can see, you know, this risk by account has been updated based off of this type of entry. So if we're seeing, you know, inventory and our cost of materials, this is in line with our expectations, but if our cost of goods sold was hitting a sales account hitting expense accounts or other balance sheet accounts that we weren't aware of. Now we have visibility into how these entries are being recorded and if they have proper controls within the organization and now we can justify, you know, if everything is being recorded the way we expected it to. Maybe we want to reduce the controls in this particular year because we're confident that the transactions are hitting the correct accounts. As we scroll over to the right, we can also see what users are entering those Cost of Goods Sold transactions. Maybe there's particular users that shouldn't be recording transactions in this area. Or maybe there's a lot of management override in this particular area. This is some things that we can gain visibility into. And now we can dig into we can see there's a bit an outlier here with 15 user 1578. Now we're able to drill down into that user and get a better understanding if the controls are operating as they should. So, as you can see, we can get a nice we can we can take a more objective approach and really quantify the type of transactions and really understand you know, the the way These entries can impact the financials. From here, we can then bolster our preliminary analytics. And the first way to do that is taking a similar approach to what we've done in the past in terms of looking at our significant variances, so it's still a crucial step because it's a nice starting point or any of these accounts that maybe had significant changes from the prior period. So right now we're looking at a horizontal analysis of our balance sheet, we can also look at it based off the income statement view. And then we can look at a vertical analysis, which is common size financials of these different accounts and how that compares to the prior period to give you more perspective on how these particular accounts compared to the prior period. So, once we have an understanding of those accounts, we can then drill down into our trending section and look into those particular accounts. So maybe we want to break down our current assets and see what that looks like over the course of the year. What we can also do which is unique to Muybridge is look at the activity. So looking at the ending balance is one thing, but looking at the activity really gives you that perspective on how the operations are being affected by these by the beat by the underlying entries, and whether the operations are consistent with prior periods. So if we get a little bit more specific, and we drill into our accounts receivable, for example, you know, we can look at we can do, we can click off a few years to to get a nice, cleaner picture and we can see, you know, in 2000, in the prior period, we had a big dip in the month of October, but it's fairly consistent in the current period. So what happened in October that didn't happen in October of this current period, and now we can document and change our approach accordingly. Lastly, you know, what we can do is we can dig into the, the different ratios that are key to the organization and highlight those time periods. The ratios do not meet our expectations. So within Muybridge, we have an algorithm called Serima, which takes an integrated moving average of the data to develop an expected range. So, from this expected range, we're able to compare that to the actual results. And if there's anything that falls outside of our expectations, we'll highlight it with a red outlier circle here. So, what that allows us to do is really target now this is days expenses in AP. So now we can target these particular accounts. And we can drill down into how these could affect this this particular month could affect the financials because it wasn't within the normal operations and within our expectations. So now we can adjust our audit plan to to really consider the timing of our day’s expenses in AP for the month of July. What's also unique to Muybridge is the ability to build out all these ratios beforehand. So we mentioned how AI can help automate a lot of those low cognitive tasks. One of the ways that we can do that is by producing these visualizations automatically for you. So, by having these here and ready to go, it's just a matter of interpreting the visualizations now, and annotating or making your comments on these visualizations. And then building out a nice work paper to support the process that you're taking within Muybridge. But once again, all this is displayed for you. So, you don't need to worry about building out these trends building out these ratios, it's just a matter of looking into how it can these, these can affect the current year audit that we're doing. So, from here, we can then go into the data table where we can drill down into those transactions and performer substantive procedures. But the focus for today is on audit planning. So, we'll dive back into our slide deck. This kind of summarizes everything that we looked at within the Muybridge platform. Right. The key is we're driving better internal and external communication through this quantitative approach. The way we're doing this is by flagging those unusual transactions and understanding how these unusual transactions can impact the financial statements. We're getting a deeper understanding the classes of transactions and how those impact our controls and the operations. And then lastly, our enhanced analytics. So these trends and ratios once we start incorporating more data, we're getting a better picture of what's actually happening within the organization. And now we're taking more specific questions to the client. So you can see how, you know the approach of the past can be enhanced through a tool like this to really develop that feedback loop in a quantitative analysis to be paired with the qualitative analysis that you've done in the past

 

Stuart Cobbe 

some of our thoughts on why this matters, as well as some examples inspired by real life situations that that kind of bring some of this into a little bit more context and why we think it's important from the perspective of driving an efficient audit. You know, the first is that I believe strongly in the value of an early warning system. And this is a great example taken from Dan Heath book upstream that talks all about how to solve problems before they happen. This is an example taken from Northwell Health, which is a health provider in the New York State area, and their emfs. Their Emergency Services Office used data used a lot of history of 911 calls to predict when and why they were happening in order to position their ambulances in various spots around the city so as to react quicker. And they ended up shaving off about a minute and a half of their response time. And you can see here that they perform significantly better than the US national average. And the perspective of Dan was that when you can foresee a problem, you have more maneuvering room to fix it. And I can think of a number of examples of when I was in audit were issues that would crop up late in the Process had this cascading effect whereby all of a sudden you had to pull in a more expensive manager, or you had to unpick a lot of the work that was done previously. I think the advantage of starting an audit with a clear data driven understanding of where problems might particularly lie within a business means it's far more likely that you'll have these kind of toppling dominoes at the end of the audit process where everybody's pulling their hair out. To give you an example, inspired by real life, this is an example of those kinds of visualizations that Michael is showing you in the tool. This is filtered for the manual adjustment, and on the left, we've got the risk burst. So, each segment of this donut represents a part of the chart of accounts that is being affected by this transaction type. The size of this segment represents the value of transactions being posted and the color represents the average risk and you can see there's a lot of Highly material movements going on in the profit and loss on foreign exchange account and these are being posted with manual journals. So, you know instantly as an auditor at the planning stage of the audit, I've got what I would consider to be quite a big red flag that needs to be waived. And I should also be a bit concerned about the control environment that's allowing somebody to post these transactions. And be I should be scheduling time to have somebody who is relatively senior and understands the complexity of foreign exchange, to be doing the work on that section of the audit because I know it might be a problem. And all of that is just reinforced by the visualization from the trending dashboard. When you look at the activity that's touching that profit or loss and foreign exchange, you see these huge swings up and down, you know, highly material movements. All of it seems to be reversed and then posted in the last period of the year. Seems pretty strange, pretty suspicious. Or possibly pretty unusual. And so, you know, the second element apart from this early warning system is really taking judgments that are based on evidence rather than on gut feel and putting engagement teams in a position where they can put that evidence down on the audit file. To use another quote from the Bridon review, regardless of the Genesis sample sizes today, oh, more to audit partners judgments than to statistical analysis. And again, this is something that I would largely agree with, from what I've seen in the audit industry. And I think what we're starting to see now is a shift towards more evidence-based justifications for sample sizes. We're seeing regulators taking a pretty dim view on kept sample sizes, for instance. So all of a sudden, the need for auditors to justify why they're doing a particular amount of work is becoming much more necessary. And so, it is An example of that closer to real life than what Michael had shown you before. This is a purchase invoices being posted, you know, it's touching largely trade accounts payable. It's going between the nominal accounts that we expect for this particular business, the vast majority of the value sits within low risk, the seasonality of the postings matches the seasonality of the business as a whole. And it's all being posted by members of the team that we expect to be posting these transactions with a low average risk. So we can see how we're building up an evidence based view of what's going on within this particular class of, of transaction. And so, you know, using both of these, all of a sudden, the planning meeting takes a slightly different slant to it. We've got client driven planning meeting, but really, we meant auditor driven planning meeting where the auditor is armed with the information, they need in order to ask the really incisive questions at the planning stage. You know, why are there so many manual adjustments made to the cash account? And this is happening much earlier in the process when the CFO and management are much more receptive to these difficult questions being asked. It also means that the preparation for that planning meeting is a team effort. It's not just the partner rocking up with their clipboard and asking questions. we're enabling other members of the team to feed into the questions that are being generated for that planning meeting. Which I think only means that the audit team as a whole is better prepared. And all of this is really being reflected in our clients. So there's this great quote from Baldwin's MindBridge's driving better conversations with our clients, which drives value the responses being you're asking better questions, which is exactly what we're trying to achieve. And that's not to say that we'll always get to the bottom of the problem in the first instance, but It certainly means that you've got a head start when you've got this kind of race against time that many audits represent. So what does it take to implement?

 

Michael Bottala 

So, what it really comes down to is as a cultural change, right, it's about altering your approach in order to take more collaborative and interdisciplinary approach to your audit plan. As Stewart just mentioned, with incorporating multiple different team members, and having them make more of an impact, in terms of the way you're approaching the audit. You know, instead of taking all these gut feelings in the experience base that we talked about earlier, we're using this data to take a more quantitative approach in order to really drive decision making. And what that allows you to do is it allows you to incorporate the younger individuals within the firm because they have the ability to interpret this data and bring value to the table and it's not all driven from the top down. You know, here's, here's the thing. Different things that have just been kind of quoted by the IAASB, you know, in regard to these different changes, so the rate of change and increasing use of big data capabilities and knowledge, you know, these are all common themes within our, our industry. And it's, you know, the time is now it's this isn't a new technology anymore. It's really something that's very prevalent and firms are taking advantage of it, such as Baldwin, CPA, missing any of these key elements in this change scorecard. It can lead to confusion; it can lead to anxiety or resistance from your team. And definitely a little bit of frustration. You know, anytime you're implementing a new process, there's going to be frustration, but the goal is to understand why you're doing it and why you're trying to get to where you're going with technology such as this. So keep in mind, you know, we want to make sure that we have a clear and definitive plan with our vision skills, incentive resources, and an action all aligned in order to really implement that change.

 

Stuart Cobbe 

To be honest, what I've what I've really seen the most is, is where the lack of resources. I mean, I think a lot of these things require an investment. And they really require people to have the dedicated time and headspace to think about what this means and whether it's coming up with the planning template or looking at how the communications need to change to the client prior to the planning meeting. There are elements that firms will need to think about and obviously we do everything we can to help, but nevertheless, the firm needs to be prepared. Thank you for coming, everybody. Hopefully you found it interesting. And feel free to get in touch with any of us here MindBridge if you've got more questions. Have a good day.