[00:01] Stella:
Welcome our panelists up to the stage, if you all can join me up here. So the first person that I’d like to welcome is Tony Saldaña, he’s the president of Transform It, a consulting organization that advises over 20 Fortune 100 companies around the world in digital transformation and global business services. Tony is the author of the famous book “Why Digital Transformations Fail.” Welcome, Tony.
Next up to the stage is Greg Leiter. He’s the senior director analyst at Gartner. Greg’s focus areas include Cloud Corps, financial management suites, financial planning and analysis, and financial close solutions. His prior roles include I.T. consulting services, business process re-engineering, managing global internal audit functions and 10 plus years in public accounting with big four accounting firms. Welcome, Greg.
Next up to the stage is Mike Gualtieri. He’s the V.P. principal analyst at Forrester. Mike’s research focuses on artificial intelligence technologies, platforms, and practices that enable technology professionals to deliver digital transformations that lead to present digital experiences and breakthrough operational efficiency. Welcome, Mike.
And then we have Bryan Degraw, you can join us on stage. Brian DeGraw is an associate principal with the Hackett Group’s intellectual property business with 25 years of business experience. Brian guides clients on their process improvement journeys through one on one discussions, published research, member webcasts, on-site briefings, and conference presentations. Welcome, Brian.
And last but not least, we have Kevin Permenter. He’s a research manager with IDC’s Enterprise Applications Team. He provides insights and intelligence across multiple areas, including enterprise resource planning, order management, financial applications, project, and portfolio management. Welcome, Kevin.
And now that we have our panelists up on stage, it’s my pleasure to welcome the moderator. Sayid Shabeer, our very own Chief Product Officer at HighRadius. All right, take it away Sayid.
[02:35] Sayid Shabeer:
Thank you, Stella. Welcome, everyone. Welcome. What a great venue to be here, right? I’m sure, I bet none of our panelists ever did a session on the field, right? So great to be here. So let’s, without much further ado, we’ll get started. Sorry about the few minutes of delay. So let’s talk in general about digital transformation first. Right. So I’ve seen some studies that say that C.F.O’s, in leading digital transformation initiatives, not only in their finance function but even company-wide. Right. From what you see in the industry and the conversations you are having because we have a great brain trust of research and experience here. What is the reality on the ground of the role of the CFO and how it is evolving? And this question, I’ll open it up to each one of you to go around the panel. So which side wants to start? All right, Tony.
Tony Saldanha:
Happy to get started. Firstly, yes, very nice to be here and welcome, everybody. A fantastic place. The role of the CFO in reality in about half the companies that I work with is not just accounting, but then also strategy. And that includes in many cases actually running business operations, running I.T, and shared services as well. Right. So it’s a recognized role of the CFO. Digital transformation is, therefore, one of the first questions that actually I work with on many of these, because, you know, the board is very clear. We’re in the midst of a revolution, the fourth industrial revolution. We know that we need to not just survive but thrive. And we know that something digital is a large part of the formula. And so, you know, the first person that would go after the CEO is the strategy, the CFO and say, what ideas do we have? And it’s amazing to me that, you know, large companies are starting to blend the resources, the money, and the people they have with incredible innovative solutions, including, by the way, some of the work that I did with you guys five years ago when I was at Procter and Gamble to basically, you know, get the best thinking of the startup and the best of the large companies. And the CFO is right in the thick of it.
Greg Leiter:
Yeah, I think in those comments, I think the other pieces that it’s just natural for a CFO to be leading a digital transformation exercise in any company. They’re typically the strategic partner of the CEO. But, you know, the key here is, you know, the CFO is a good individual, helping drive change into the organization. And certainly, with digital, you’re going to have to change your skill sets with your employees, particularly from going from, you know, the traditional finance roles of, let’s say, continuous improvement and, you know, analysis of results and so on. Now they have to be digital interpreters, be able to translate into business results. That’s just a natural place for finance to be. And certainly for the CFO to help drive it, right?
Mike Gualtieri:
Yes. So we have some recent data, our survey data that shows that digital transformation, 60 percent of the companies say that A.I. is a critical component of their digital transformation. So I will cover Artificial Intelligence. I’m coming from that perspective. But if Artificial Intelligence is linked to digital transformations, it begs the question, what does the executive team, what a CFO needs to know about that? Is that different? I mean, fundamentally, artificial intelligence is software, but it’s a little bit different because it’s non-deterministic, it’s probabilistic. So that requires different funding models because you may find a use case for a machine learning model. You know, it’s not like regular software. You don’t necessarily know that it’s going to work until you try it, right?
So the funding model has to be different. It can’t be: Here’s one use case. What is the return on that use case? What is the funding requirement? You typically have to think more like a VC who does their due diligence, invests in 10 or 12 companies, and believes every single one is going to be successful, but statistically only a few. So that’s a very different way of thinking about how you have to fund A.I. projects. Now, having said that, that’s more about the building like using data science teams. Increasingly, there’s the buy option. Almost every single ISV I talk to is building machine learning or some sort of A.I. capability into the application. So I think the CFO also has a role to ask those questions. Why are we doing this? Should we be doing this now and in three years will an application already have that capability?
Bryan Degraw:
You know, what I think is interesting is the whole concept of transformation. Ten years ago, if you said, hey, we’re going through a transformation. What’s the project? I think what we’re seeing today is, if you’re not transforming in some way, you’re behind. The transformation today is almost replacing continuous improvement. You’re constantly transforming. And we’re the CFO sets. I mean, we do a key issue study every year. The top three things that the enterprise says we need help from finance with is, is cost efficiency, growing? The organizations have growth strategies. And then also enabling skills around analytics. All three of those, if you’re not transforming or improving your skill set around those areas, you’re behind. And again, to the point that the CFO sits in a perfect position to look across the enterprise and say, where are our gaps in those opportunities?
[08:44] Kevin Permenter:
One of the things that we’re seeing, of course, is digital transformation. Instead of it being sort of this amorphous kind of top-line task that is taken on by the CFO. Maybe the other folks in the C-suite, what we’re seeing is that it’s driving down to the use case level. Right. So, in fact, you know, some of the things that I saw at the Hackett Group presentation, the idea of, you know, driving automation toward deductions or disputes or collections. That’s what digital transformation looks like. If you look at it from the use case perspective and more of a ground grassroots kind of approach, then you can get your head around the concept of digital transformation, a much easier fashion.
[09:32] Sayid Shabeer:
Right. So you guys just turn up on a couple of things, right. Words when we think about digital transformation. You brought it up from the ground up, right. But also from the top down, not in the media. You do hear a lot of buzzwords, machine learning, cloud, RPA, mobile, you name it, right? That is associated with digital transformation. So in your assessment and again, I’ll ask you for the top one. Right. In your assessment. What’s the top one? Maybe the top two, one thing initiated is finance function should focus on in 2020 in their digital transformation journey to make the most impact for the business. We know that’s key. Right. And again, in spite of the title of the panel, you don’t have to say. Alright. You picked the right one. I mean, what is the top thing that the audience here should be focusing on or thinking about when thinking about digital transformation?
[10:31] Kevin Permenter:
Right. We’re all here. I think the first pass at it and I won’t say so, but we’re all here because of the sort of core aspect. And that’s, you know, cash. Where’s my cash? You know, how much do I have now? How is it being used? So what we’re seeing is a lot of A.I. machine learning going toward the collections process, going toward some of the treasury processes, going towards some of the cash forecasting processes, being able to apply machine learning in A.I. and even some of the more advanced topics like cognitive computing and neural networks to the cash problem. That’s where we see things going.
Sayid Shabeer:
Right. What’s your top one?
Bryan Degraw:
I think the top one would definitely be the analytics aspect of it. Obviously, you know, to your point. I mean, we really want to be more predictive and really know that our process is stable and there’s no surprises and the ability to harness technology to help, you know, where you are today, where you’re going to be tomorrow. I mean, I think that’s really powerful. And again, a spot where the finance and CFO, I mean, that’s a key service that their partners and the business are asking them to provide.
Mike Gualtieri:
So I just want to read some data here about digital. We asked large enterprises what their top business reasons for doing digital transformation are. These aren’t necessarily surprising, but I think there’s a tie in to a high number one choice. Sixty-one percent better customer experience. Number two, increased efficiency. Forty-nine percent innovate products and services, 36 percent. Now, the reason why I think what’s interesting is that when you think about A.I., many executive teams think it’s just about innovation and new business models. And it can be. But our A.I. surveys show that it’s in total alignment. They want to use A.I. for the most part to do customer experience. When we also asked what is the greatest challenge to digital transformation, security, technology, strategy, data issues with culture- a distant fourth now when thinking about artificial intelligence.
It’s data technology. It requires data. It’s driven by data. So I think one of the things that we put out, this prediction, is reporting A.I. and we describe what we call data crybabies. Because a lot of people were saying, well, I can’t do machine learning. Data scientists spend too much time getting the data. But when we actually talk to companies who were successful, it was always an executive that stepped in and said, get them the data. And low and behold, they had the data they needed to get the job done. So we think the CFO can play a role in helping companies invest in the data management technology that they need to get that data done. And the second thing is to work on consequential use cases. Right. Not ‘toy’ use cases that someone just dreamed up with a guy, but a list of consequential priorities, business use cases. There are not too many people better to look independently at those use cases and make that determination than the CFO.
Greg Leiter:
Great. Planning and forecasting is one area I think is going to be a really interesting one for this year that the markets for the cloud have been, vendors have been growing like 30-40 percent a year. You know, the vendors I speak to her grow leaps and bounds. Probably half the clients I interact with on the planning side of questions are still using incredible amounts of excel in their planning processes. And these solutions out in the marketplace right now have done a really credible job and making it, you know, the ease of use, implementation, putting guardrails around your planning processes.
And to me, the other thing here is the vendors are also putting on a lot of interesting A.I. into these tools now with predictive analytics capabilities and moving bit by bit to more prescriptive analytics. You know, now I give me indications of what my forecast will be, but you know what I should do when someone says they have that information. So it’s a very interesting area right now. I think it’s going to be one that’s important. Now we need to know what happens in terms of closing your books and reporting results. But you need to know what’s going to happen in the future. So that’s why I think planning forecasting will be a big one.
Sayid Shabeer:
Great. And Tony, you’ve been a practitioner besides being into consulting now. What do you see in 2020 as a top thing that’s on top of our folk’s minds?
Tony Saldanha:
So my recommendation to the CFO is we’d actually be neither in terms of technology nor in terms of a work process because every business is different, the situation’s different, and their priorities are going to be different. So the goal I would set for CFO would be tied to your business model, identify the top five material transformational ideas, and in the next year make one of them completely successful in a significant part of your business. So I need to unpack that a little bit. You know, as my colleagues have said, the challenge here is we kind of get caught up with buzzwords, technology, you know, and so on and so forth. Real digital transformation is how you rewire your organization to succeed in essentially the fourth industrial revolution. And we know 90 percent of that is organization change management and maybe only 10 percent of that is technology. Right. And so when we train, when we chase after a technology or a work process, we lose the linkage to our owners, you know, either personal owners or governments or Wall Street. Right. And so the job of the finance manager is to start with total shareholder return if that’s what you know. If you’re a public company, figure out, as I said, the most material use cases in your organization. It could be cash. It could be something related to the supply chain. It could be something related. In the work that I did at Procter and Gamble to drive disruption. If the prize wasn’t $50 million, that was not interesting. Right. So pick something that’s material. And then secondly, as I said, go from 5 to 1. Assume that out of the 5 big ideas. It’s OK to fail on four of them because if you get one 50 million dollar idea, you’re a big hero. Right. So that’s the way I would set the goal.
Sayid Shabeer:
That’s a great insight. Right. In terms of. You know, it’s always a portfolio play. Sometimes you’re doing different things, but you get that one chance, and then it spreads, that you can do more of it. And the confidence builds. Let’s shift a little bit towards the challenges of AI. We’ve heard a lot of exciting things, changes, and the business benefits that can come from it. But, you know, kind of thinking about the challenges or risks. So let’s start with you, Tony. The availability of high-quality data that was already mentioned, as well as the security privacy around data. So you don’t have to give any proprietary information. But how did, in your experience speaking to you, how did you address the data-related issues that significantly impact any air-based solution implemented?
Tony Saldanha:
Data is about 90 percent of the challenge in any A.I. game. As I’m sure you guys have data to back up that, right. But the way we address this at Procter and Gamble is first and foremost to, again, build on something where the palace had earlier started with a use case. The worst mistake that you can make as the data architect or the business intelligence leader and I ran some of those roles at Procter and Gamble is to basically go off to world hunger.
You know, we’re going to harmonize all the financial data and then we’ll build use cases. Bad idea. Definitely. You never do that. Right. So start with a use case. And so let’s say we’re going after deductions and claims management, which was the real use case that we actually went in and built the solution co-develop between Procter and Gamble and HighRadius. And this was the use case whereas you get accounts, receivables claims from vendors, Walmart, our Target, so on and so forth. Instead of having hundreds of people across the world look at these claims and say, is this valid or invalid? You know, could you not have algorithms that could make that same determination? And, you know, eventually, we succeeded. And, you know, that’s built into your product now, but then bringing together that data. If we had started from- OK, we’re gonna take the next year to build all of that data and harmonize it and then feed the A.I. algorithms, you would be dead. So the approach that you take is you basically do things in parallel. You essentially start with the data that’s available. You train the models knowing that this is incomplete data. But you test the hypothesis to say, if I had more of this data, it would become more accurate. That builds momentum. You’re able to say in the glide path of accuracy, I would like the algorithm to be 99 percent accurate, but with the data in my glide path, I needed to be 80 percent accurate. And I’m going to implement it. So that’s point one.
The second thing you do is you don’t think grand about I’m going to replace these 50 people with this algorithm. You start with machine-aided beliefs invests in decision-making. So I’m going to have these people and these algorithms work together. And then algorithms get smarter and smarter. So when you take a more pragmatic approach like that, you build up confidence and you build up momentum in the organization.
Sayid Shabeer:
Great. Thanks. So this is for Mike. You know, when we think about the Fortune Thousand, right, especially about A.I., data skills shortage often comes up. Right. So what recommendations do you have for organizations that have this challenge in the short term and the long run?
Mike Gualtieri:
Sure. So two things in the short term. And one is that I think it’s a myth that data scientists are unicorns, that a data scientist has to know the math and know the business. They don’t have to know the business. They have to definitely team with someone who knows the business, right? Because a data scientist there, their approach to building a model is pure math at some point. But they need to be able to hypothesize about what data will actually build a good model. And who better to hypothesize what variables may matter than people in the business, you know, who know the business.
So I think a lot of companies have been chasing that unicorn version of the data scientists. And it’s not necessary. Nice to have. Yes, it’s not actually necessary. The second thing is there’s been some amazing new big because of the popularity of AI and machine learning, there are some vendors who have been responding to the market. The research community is doing a lot of work on tools, collaboration tools, data scientist data science teams are like where software development teams were 20 years ago, not working very smoothly. They didn’t really have collaborative tools. That’s changing. So they are upgrading the tools to more collaborative tools and the most exciting technology is auto ML and auto ML, automatic machine learning. It’s not fully automatic, but this can dramatically improve the productivity of data science teams upwards of a thousand percent. And here’s why. Because when a data scientist is building a model, they hypothesize about some data and then they apply one or more machine learning algorithms to that data. And it’s sort of a very iterative cycle. Auto ML sort of allows them to sort of set this whole thing up and then iterate a thousand times, whereas they would have to do it manually before. So those same auto ML tools in some cases can also empower non-data scientists, sometimes called citizen data scientists, to participate. We don’t believe that a citizen data scientist, someone who doesn’t know how to statistically evaluate a model, should be creating machine learning risk models, making million-dollar loan decisions. But there are some other consequential use cases that they can’t do. So there’s a lot happening to make teams more productive in general.
Sayid Shabeer:
And I might add if you’re hearing terms like auto ML and you’re wondering how the heck am I going to get this in my organization? Don’t worry. You know, a lot of those things that were just talked about, HighRadius, is this, you know, all the heavy lifting at our end about machine learning is the stuff that we’re doing so that we can actually bring it to the corporate’s, right.
Bryan, so when thinking about costs, benefits, right? Evaluations for the organization are often ‘requests’, especially if you’re thinking about machine learning and/or even other technologies, requests probably at another level of sophistication, judgment, experience. So what are some of the best practices that you would suggest for someone that is embarking on an online digital transformation initiative? And we got some pointers from Tony around that now. But anything, from your experience talking to corporates that folks should keep in mind when they are evaluating cost, benefit, ROI, and best practices around that?
Bryan Degraw:
Sure. Well, you know, here in this conversation and thinking back to my career and maybe in some of yours, it almost sounds similar to when we’re looking at ERPs. And, you know, the promise of what those solutions are going to provide. And, you know, a big piece of it that we always talk about is you go slow, be methodical, look at it from a process perspective. You know, we share, especially around technology and automation, as the tools are as powerful today as they were back then, you know, as they evolved, they could definitely automate best practices, but they could equally automate bad practices. And so really evaluating what you need to do? So, you know, in terms of what we suggest is in terms of evaluating these, you know, machine learning and A.I. is really deconstructing your work. And if we’re focusing on financed practices and processes, specifically- customer to cash, he’ll decompose that into what is structured work, what is knowledge work and then interactive work, interactive meaning, you know, either it’s collaborative within your organization or obviously very important collaborative with your customers. And then break it down into a heat map and from a heat map, say, OK, where are our opportunities? You can’t do them all. And then the big thing is the customer. What is the appetite?
You know, I think we all have experienced it. I have experiences where I want to use a chatbot, but I may be multitasking. I’m on a call with my team, but I’ve got to do something personal for the home or something. That’s great. But there are also times sometimes where I need to talk to somebody. So the concept of thinking about this ecosystem of solutions, especially with the customer in mind, is that it’s not one size fits all. And you really have to think omnichannel. You know, I know there’s a lot of debate between what is multi-channel versus omnichannel, but omnichannel is regardless of how the customer wants to interact with you. You can provide the same experience.
Sayid Shabeer:
Yeah. Greg, we often hear about, you know, with the proliferation of A.I. based solutions, including in the finance function, questions about if at all there are any ethical issues in the B2B world. Right. Of course in the B2C world, there are things like consumer credit and so on. We have heard about the B2B world. Are there any that we should be concerned about, thinking about?
Greg Leiter
Well, you know, ethics is a very good question, because I think there’s a lot of concern and fear about A.I. as a black box kind of a side to it. And I think what Mike was alluding to, it’s been the domain of the data sciences, more or less in the backroom as opposed to being out front. And, you know, Gardner takes the view that you should be having a governance process over your A.I. initiatives and take more of a glass box approach to it. You have more trust, transparency, and also diversity because you’re making decisions using some like, although I know that maybe you have biases against, you know, certain races or gender or things like that. In terms of like, extending credit to somebody, for example, again, that’s, you know, more of the B2C side.
You know, certainly, on the B2B side, there’s gonna be less of those issues. However, you know, if I was an employee in an organization that was running lots of A.I. now. I mean, am I going to trust what’s telling me this is going to impact my job? So there’s that openness that you as an organization need to have with your employees. Gardner also predicts that at some point, probably, in the next five years, it’s going to be someone they trust actually between two companies because their algorithms have done something that caused any competitive situation. So, you know, the imagination can get pretty wild. I don’t see much on the B2B side necessarily, but you can imagine things happening that could be pretty negative.
Sayid Shabeer:
So where in finance do you see the most A.I. adoption? You know, just throwing things like payments, credit management, what have you. I wanted you to mention just one area where you think an A.I. adoption could be high.
Kevin Permenter:
Well, so whenever I’m talking with users, kind of like yourself or in the audience, the areas that sort of rise to the top in terms of pain are collections. So, the process of automating a lot of the contacts, the emails that go out, making sure that the right email goes to the vendor goes at the right time. And being more proactive, I saw some of that stuff in your presentation. So, there’s a real need for, we see a real need to automate that area. I’ve seen a lot of companies, a lot of your competitors, taking a hard look at how to automate and use it to address that pain.
Bryan Degraw:
I would definitely echo that, it is strategic, in terms of how to touch customers, looking at it from a collections perspective. You know, machine learning is being used in the cash application process, coupling the remittance advice along with the payment. Again, focusing specifically on customer-cash, dispute management is a big area where the tool could help to be more predictive, interactive. I think we have a big opportunity.
Mike Gualtieri:
I guess what I would say is that, if you’re not an A.I. or ML expert, you’re not quite sure what it is. It completely translates into your mind- prediction. What can I predict? Because any of these applications is prediction, predicting default, predicting who will be less likely to complain if you don’t pay them. This is how you find in use cases, that is consequential if you just ask yourself the question- What could I predict to make this process better and include better business outcome. And in that way, you don’t freeze when people are talking about A. I and ML. Because 90 percent of it is a model that makes a prediction about something.
Greg Leiter:
I have to agree with that, my area was to predict the value of the forecast, that’s the best use case right now.
Tony Saldanha:
I have to talk about deductions management that we guys did together. For two reasons, one is that it’s available. So, you don’t have to actually build it, it’s now a part of your tool. In large companies, claims and deductions are a multi-billion dollar type of problem and you know, 50 percent of the claims that come in are valid and the rest is invalid. In terms of dollar value, you give away 90 percent and then deny 10 percent. There is a huge opportunity there, and again, the product’s availability.
Sayid Shabeer:
Great. Thanks a lot for the time. A lot of great insights, hopefully, you guys, as you think about A.I or about digital transformation more generally. Hopefully, you can not only hear their advice in terms of the experience that they bring but also take a look at a lot of the technology. We’ve worked with clients, a lot of you, to build these use cases, and to solve these problems for you. And how that fits into the broader umbrella of digital transformation, it’s key. Whether you are advocating for your C.F.O. for this transformation, in supporting yourself, to drive digital transformation in your company, we’re here to support you as HighRadius. And hopefully, in that journey, these folks here gave a lot of insights on taking that journey. Thanks a lot, guys. Thanks.
