Transcription: #26 - Jonathan Chizick: The Intersection of Outsourcing & AI
#26 - Jonathan Chizick: The Intersection of Outsourcing & AI


Aderson Oliveira: I've spoken with Jonathan Chizick about the connection between outsourcing and AI. We're talking about AppZen and how its AI-powered solution comes to help businesses automating back-office tasks like expense report creation, auditing, and compliance as well. In the end, he said that humans and their jobs are not going anywhere soon, but AI will definitely be part of the future of the world.

Hello, hello, Aderson Oliveira here. This is another interview for the OuchSourcing Podcast where I talk to business owners, to specialist, to experts about outsourcing and the trends that goes together with outsourcing or that may affect outsourcing, like what we'll be talking about today. Today, I have with me Jonathan Chizick. He is the Head of Marketing at AppZen, which is an AI-powered solution that automates expense report creation, compliance, and auditing as well. Jonathan, welcome.

Jonathan Chizick: Welcome, and thank you for having me. Good to be here.

Aderson: Perfect, very good, Jonathan. Let me start with this. Yesterday, Jonathan, I was talking to a good friend of mine, and I told him that I was going to talk today with a specialist about AI and outsourcing, and he just looked at me, he laughed a little bit, and said, "How are you going to connect outsourcing with AI? Is there really a connection?" That's my first question to you, Jonathan. What is the connection between AI and outsourcing?

Jonathan: If you think about some of the reasons why outsourcing took place was because there's task-oriented activities that were done that can be pretty well defined and could be easily outsourced to someone who could specialize in just repeating their process and not necessarily have to know more about the whole context of the organization, or the business, or that sort of thing. It was easy to outsource to lower costs, typically it was because lower cost labor countries. You can outsource those processes.

What we're doing with AI is taking those task-oriented activities, and a machine can do those right. They're very routine, they have different rules, machines have capabilities to learn and build on those rules, and basically replicate what humans are doing in these outsourcing activities. You kind of think of it as if I was outsourcing to people, you could be outsourcing to machines or a positive spin on that would be that it's making the people who are currently doing these tasks much more effective.

Aderson: You probably have a point which I have actually written that up here is instead of outsourcing to human intelligence, we are outsourcing to an artificial intelligence, that's it. But, now let's bring it back to AppZen, how does that play with AppZen?

Jonathan: We're an AI-powered platform for enterprise back-office automation. As you mentioned in the intro, our first product is a tool that automates T&E expense report audit and does that in real time. We're not removing people completely from the equation because what we're doing is enabling people who are currently involved in these activities. We're looking through all this expense reports, going through receipts, verifying that, "Okay, this merchant is exactly what the employee said it is. These amounts match. They didn't accidentally type a nine instead of a six, and then is this an establishment that they are allowed to go to, or does this violate some sort of compliance policy? Are they having dinner with a politician or someone?" It could be totally mistaken and accidental.

What we're doing is identifying these risks and highlighting them for the people who are still involved in that audit process. So, now these auditors are able to audit all of the expenses as opposed to random sampling or just trying to guess as to which ones or just high-value ones that they want to look at and get to a deeper level of detail. Now, they're able to spend their time instead of checking receipts, 95% of which are going to be okay, they can spend their time focus on, "Hey, this one's wrong. What do I do with it within the context of my business and my business's policies?"

Aderson: Does that become a bit of a no-brainer for you to present that to a company, and it's a no-brainer to adopt that? Does that become a very easy sale for you guys?

Jonathan: Sales are never easy, but we are getting a really good reception to it. It's one of those where people get it and they're like, "Okay, now I just need to see that it actually works and can do what you say it does," and then compare it to their existing processes that they have in place, which typically have been in place for a while, like they're used to them and know how they perform. From that standpoint, like we'll move on proof of concepts, we'll prove out the ROI and show that we actually do what we say we can and that's usually where people get really excited. Now, it is kind of a no-brainer at that point. There's got to be some extenuating circumstance why you'd say no at that point.

Aderson: Let's talk a little bit about the typical client profile of AppZen. I went through your interview with my friend, Justin Jones, and you guys mentioned a little about that. But, talk a little bit about the client profile of AppZen.

Jonathan: Our folks are mid or large enterprise customers, and really, our tool is valuable to any size company. Just what we found is that these larger companies, typically, they know the pain point. They have processes in place, their resource is doing some level of audit. Every company has to process T&E, but larger companies have to do it in higher volumes, so they typically have a function that's dedicated to work on a function that's specifically responsible for it. They understand that the ROI of what we do and they can clearly see that the benefit of just adopting AppZen. Like I said, our application is valid for any size. Sole proprietor could use this just to keep track of their receipts and make sure that what they're submitting to the tax authorities is correct, but that's not really where we're focused in selling to right now.

Aderson: You mentioned, at the beginning, that this is your first app or your first offering. Does that imply that there are more offerings there on your bag of offerings at AppZen?

Jonathan: Yes. If you think about it, the technology we're using, the same thought processes that were taken from humans can be used in different back-office procedures today. What we're focusing on right now is T&E expense audit, but in back-office accounts payable, invoice processing is a huge undertaking for most businesses and there's a lot of money falling through there. You could see that the same sort of things we're looking at in expense reports could be applied to areas like looking at invoices or the accounts payable area.

What we found is, really, the more you can say focus on are limited to vertical area with your AI, the deeper you can go and the more intelligence it can provide, and more value it can provide for the end customer. So, rather than having a generic AI platform that kind of cuts across surface level across many areas, it doesn't really provide as much value to specific groups within the company. So, we decided to go deep with T&E, provide a lot of value of looking into the future on how we can extend that technology to other areas.

Aderson: What you’re referring there is narrow AI as opposed to -- is that wide?

Jonathan: Yeah. Instead of saying, "Hey, this is an AI for compliance, and it looks at everything from employee relationships with vendors to where do people eat, and are other people in the same -- it's looking at, specifically, one function that a business is doing and getting very deep into how we can look at things and how we analyze different expenses and other things within that one niche.

So, as opposed to an analogy that might be -- it makes it easier to stand it up, say, with marketing. You have an AI tool for marketing. There's a lot of AI tools that say they help with marketing. This will be more along the lines of, "Okay, we have an AI tool that's going to automate your email marketing," and then we'll take, once we become an expert there, we'll expand to the rest of your digital marketing and social media, and maybe some other areas of marketing as opposed to just an AI tool that cuts across a surface level of all marketing.

You can also think of it as if you come to your company, and the company says, "I'm going to automate all of my back office." That sounds great, but that would be a massive undertaking that would take a very long time to convince an enterprise to adopt that just because the risk that's involved, and there's so many different processes. We just found it's better to get a foothold in one area where you can show some clear value, clear ROI, kind of go from there.

Aderson: Let's talk a little bit about the kind of results, Jonathan, that you have been able, so far, to bring to AppZen clients. Without mentioning names, can you maybe talk a little bit about we were able to do this in this amount of time, reduce the cost, improve the delivery, the speed of this, can you talk a little bit about that?

Jonathan: Sure. Across all of our clients right now, we're averaging about 10 times more effectiveness, and effectiveness measured in violations whether it's accidental or intentional, but things that violate a policy, or duplicate receipts, or if there's some sort of error in the receipt we're identifying 10 times more than the current manual processes at the same time reducing the workload required for the business to identify that. We're also seeing that that's driving a hard T&E cost reductions. So, right up front, companies are not paying for things that they shouldn't have been paying for before. They just didn't know that a duplicate receipt was going through, or there's personal expenses that were just flowing through that, unless you look at the receipt and you had a technology to actually read the receipt, you'd never know that someone was expensing a necklace that they brought at a jewelry store as much.

A specific example, we got a recent client who they had an audit team of few people that was managing in the back-office and the volume of expense reports because the business was growing and growing. They were not getting extra budget to expand the resources to monitor those. They were looking at how we keep process those expense reports in the quickness that we've committed to. Because, employees submit their expense reports, they want to get paid right away. We can't keep delaying this, and that's what's going to happen. We're going to have to go to a week, or two weeks, or longer to process these expense reports. Once they implement it absent, they've gone down to -- they'll actually be able to shift some of those audit resources to other areas, and then at the same time, process most of the expense reports the same day.

Aderson: Jonathan, I'm not super familiarized with the outsourcing aspect that goes on with report auditing like that, like where AppZen comes in place. So, let me ask you this. How big is the outsourcing component of that indigenous to you, or companies in general do that more in-house? Again, you see a percentage of the market there that they outsource that component a lot that you are bringing to the table.

Jonathan: There's a mix. We take companies, say North American-headquartered companies, a very small percentage of their audit is done in North America. Now, some of that is outsourced, some of that is done in shared services centers that are still within the company, but they're typically located in the Philippines, or India, Eastern Europe, lower wage areas of the world. So, then this is like outsourcing, but it's going to be outsourcing within the company, and to these decentralized shared services areas. That's why we're focus on mid to large enterprises, because by the time companies get to that size, they realize that there's some significant process efficiency to be had by centralizing those activities. In the smaller companies, sometimes the audit is done, if it's done at all, office-by-office just by the admin, or HR, or whoever happens to kind of get stuck with that job.

Aderson: It's funny because offshoring has taken jobs from North America to overseas, and then it seems to me that AI will start to take jobs from offshoring to computers, in general, just to code. Do you see that across the board? Because, you are in the Silicon Valley, you are in San Francisco, so do you see that happening more and more? What are some of the trends that you see towards that?

Jonathan: It's the early stage. Seeing the early beginnings of that sort of transition and seeing even the -- you take these outsourcing companies, very large BPO shops that have been taking on outsourced business from large companies for 15 years now, they're very interested in tools like AppZen from the aspect of -- we know that we need to get more efficient, because if we don't get more efficient, we're out of places to go to chase who can we pay less to and move everything to that country. There's political risk, all sorts of things going on that makes that not such a great idea as opposed to, "Hey, if we can implement these tools within our existing processes, we can become more efficient, and then maybe not necessarily -- the companies might not insource those previously outsourced processes, but now they're getting even better value from these outsourced companies.

Aderson: What you're saying is that you don't -- because my initial thought is that AI will come to replace them because of cost, as well, cost being one of the big drivers, efficiency being the other big one. You see them more working hand-in-hand, not so much in place each other?

Jonathan: It could be both. It depends on the situation. We see that working hand-in-hand from the standpoint of if you have -- take a North American-headquartered company, you have an accounts payable group, or some finance group that once you go up all the chain of those dollars that have been flowing through T&E expenses end up on their P&L. Now, maybe they're able to take on more of what had been outsourced before, because they're not having to do -- the reason they outsource these processes where they were very manual, repetitive, and not high value add, if we take those same tasks and have machines do that where it can do millions of those tasks very, very quickly at very low cost, now that the people who are still there at the company are able to just look at the results and decide, "What do I do with this?" and just spend more time making business decisions based on that data.

Aderson: One of the things that I have been seeing lately, and actually I have come to experience that is something very simple, a very single point of AI technology there which is x.AI, or even, which is one of the tasks that virtual assistants, or assistants in general, do quite a lot is setting up appointments, and those platforms are taking them away from those jobs and automating them.

Again, the trend is there. What would you say for outsourcing companies that focus on the lower-value type of repetitive activity? Because, they must be seeing that coming. How should they get ready for that wave of automation and seeing themselves replaced, at some point, with AI and technology, in general? What would you tell someone that drives a business like that, outsourcing low value type of activities?

Jonathan: You may have answered some of it right there. I think they just need to look at where the landscape is going and figure out how they adopt these tools. When you look at what goes on in those outsourced offices around the world, there's a lot of automated tools in place. There's some RPA tools, which kind of glorified Excel macros, but there's technology in place that helps those employees do these manual tasks faster and more accurately. AIs are kind of taking that to another level. They'd certainly be looking at where are these animated tools going and, "Which one should I pick to make my existing workforce much more efficient?" because if I don't, someone else is going to do that, or companies can adopt these tools internally inside, "I don't need the outsourcer anymore, because I have machines that can do it for me at the same price or lower."

Aderson: I guess that one way that you could see that, Jonathan, correct me if I'm wrong, but is maybe trying to shift a little bit more from the labor-intensive type of work to the more creative intensive type of work, if that can be done overseas, offshoring with outsourcing. I think that a shift like that, because by the time that we get to a point that AI is doing a lot of the creative as well, there's still a gap to go that far. So, I think that moving from labor-intensive to more creative-intensive type of work, don't you think so?

Jonathan: We do see some of that in our business which would be for these outsourcing companies or these shared services areas, they should be looking at the workforce that they have. Do they have the skillsets that we're going to need them to have in the future? I think you definitely highlighted a good point, as opposed to, say, today an auditor, I'll take the T&E example. Their task is to look at the receipt, and match up numbers, and match up a name, and then click a button saying, "It's good, or if it's not good, pick one of five of different responses to say why it's not good." That can all be automated very, very easily.

But, the real value add comes in the ones that aren't okay. What do you actually do with it? Does it need to go to HR, does it need to go back to the employee, does the manager need to be notified? What happens to that? And that's typically not a skillset that these outsourced companies have today, but it could be a service that they provide in the future.

Aderson: Let's talk a little bit more -- going back here, I'm jumping all over the place here and there, but let's jump back here to AppZen. How quick, I would say, a new company, a new client that engages with AppZen, how quickly can they get to start to see some results and start to see some - actually, forget about results for a second - start to see some implementation done within AppZen connecting to their infrastructure?

Jonathan: Absolutely. If they're working with Oracle Expense or Concur. We have, even, direct API integrations with them. We're SaaS company, so with them, the mission starts right away. Once the authorizations are done on both ends, it's turned on. We'll start seeing results. Now, we take about anywhere from two to four weeks of this calibration to ramp up where we're running in parallel processes, the existing process, and then also apps in, and that's to help train the tool. We have best practice rules built in that we'll start with, but then every company has their own idiosyncrasies of what's going on with their T&E, and this is the time when a tool learns. Within a month, they're kind of steady state and false positives are very low, and they'll start to see some significant benefits in the efficiency of their processes.

Aderson: So, train the tool to work within that client-specific policies and procedures, correct?

Jonathan: Yeah, exactly.

Aderson: Perfect, and you said that, during that month, there's some fine-tuning that needs to be done. So, even though it's an AI tool, there is still a lot of human creation, at least, at the very beginning to fine-tune things?

Jonathan: A lot of it's already been gone. We have enough clients where the tool that was implemented typically performed pretty well right off the bat. But, what we'd like to do is make sure that once they're relying on AppZen as their sole process to audit that they can be confident that we're delivering accurate results. So, we'll run these two in parallel, and then anything that pops up as exceptions where we have their existing auditors saying, "Well, wait a minute, that's okay. That shouldn't be flagged," then we can put that back in the tool and the tool won't flag that again. Sometimes, there might be something that comes up. "Hey, why did the tool not catch this one thing?" and then we'd go in and make sure the tool is trained to catch that going forward. Typically, that's two to four weeks, but sometimes it's right off the bat running really well. Other times, just takes a little bit longer.

Aderson: What is the long-term vision for AppZen, Jonathan? In five years, where do you envision the company and the service to be in? Talk a little bit about the future here.

Jonathan: We see AppZen as being this platform that automates the business back office. You can think of all the processes from expense reporting to accounts payable, managing invoices, potentially even going into some of the accounting areas, like closing books and creating PNLs. If you think of every back-office function that's done today, there's some level of automation that could be applied towards it. We have some patents on our technology, and we're going to continue to look as to where the next area is that is most right for applying that technology and go from there.

Aderson: I knew that there was a link between AI and outsourcing as well, and again, I'm hoping that this will be the first of many AI conversations that I'll be having. Jonathan, anything that we have not discussed here that might be relevant, whether it's about AppZen, or tying that back to outsourcing, anything that we haven't spoken about that is relevant?

Jonathan: I think we hit on all the key points. If you're an outsourcer, you are probably already aware of AI, but definitely it's something, it's a trend you need to stay on top of.

Aderson: I usually ask my guest for the one thing that they would like people to leave this conversation knowing about, and now I'm better because I've watched this 20 minutes, 25 minutes, and this is the one thing that I'm going to leave with. What is it that you want to highlight that people leave this conversation knowing about?

Jonathan: That's humans aren't going away tomorrow, but as you're looking at what the future of work is, AI is certainly going to be a part of that, and if you're worried about your job, just think about what aspects of it would be very, very difficult to automate, or couldn't be automated, and maybe those are the skillsets that you need to focus on. Or, if you're an outsourcer, you need to focus your workforce on having that skillset. It's going to be kind of a long progression for all business to transition, but I think there's a lot more positives in putting AI into businesses than there are negatives.

Aderson: Perfect, very good, Jonathan. I would assume that if people want to reach out, maybe, to you or maybe to AppZen, they can go to, and I guess reach out. There's a Contact Us or something like that.

Jonathan: Exactly., or you could send us an email at or I'm at, pretty easy to find, got a LinkedIn too, so feel free to reach out any way.

Aderson: All the links that were mentioned, including the email addresses, here will be posted in the show notes. Jonathan, thank you very much for taking the time from your busy day, and I really appreciate the knowledge sharing that you have provided here, a lot of knowledge about AI in connecting the dots for us with outsourcing as well. Thank you very much for you time. Talk to you soon. Thank you, bye.

Jonathan: You're welcome. Bye.

I'm an Outsourcerer. I'm a DNN Geek. I help people with their sites @ DeskPal. I'm a #Pomodoro practitioner. I'm a husband and a father of 2 beautiful girls.

Read more posts by Aderson

Download the Top 50 Outsourcing Sites

Choose the right Outsourcing Site

Don't start your new PROJECT before you download the Top 50 Outsourcing Sites guide! It will save you time and help you choosing the right site.


Aderson Oliveira
Aderson Oliveira