-
Subscription billing
Manage even the most complex billing models with ease
Subscription management
Simply your customer lifecycle management, no matter how you bill
Revenue recognition
Automate your revenue recognition schedules to ensure GAAP compliance
Metrics & analytics
Leverage your billing, invoicing, and customer data for deep insights
By Go-to-Market Model
By Function
-
Subscription billing
-
Subscription management
-
Revenue recognition
-
Metrics & analytics
- Take a tour
By Go-to-Market Model
By Function
-
-
Maxio Platform
Your financial operations platform to manage subscriptions, billing, revenue recognition, and financial reporting.
Featured Modules
Advanced Billing
Automated recurring billing to power your product-led or self-service strategy
A/R Management
Reduce your A/R balance, drive down DSO, and get cash in the door faster
Advanced Revenue Management
GAAP/IFRS-compliant revenue recognition your auditors will love
Milestone-Based Projects
Bill customers and recognize revenue based on future events (i.e. implementations)
Expense Amortization
Expense accounting for prepaid expenses, fixed assets, and sales commissions
-
Pricing
-
Content Library
Get practical tips and tricks to improve your daily financial workflows
Blog
Level up your billing and finance knowledge, and keep up with the latest in SaaS
Podcast
Listen to engaging conversations with leading SaaS and finance experts
Maxio Institute
Get the latest insights into SaaS growth based on real companies’ billing data
Case Studies
Read real customer stories, and learn why B2B leaders love Maxio
SaaSpedia
The encyclopedia of SaaS finance terms and metrics
-
Quickstart guides
Everything your team needs to get up and running with Maxio
API documentation
Integrate with many environments and programming languages via our REST API
SDK code libraries
Available for JavaScript, Python, Java, Ruby, PHP, TypeScript, and .Net
Webhook documentation
Send HTTP POST requests for easy parsing in almost any programming language
-
-
Maxio Platform
-
Get a Demo
Revolutionizing FP&A: How AI is Shaping Financial Analysis with Nicholas Boucher
July 11, 2024
Episode details
This week on the Expert Voices podcast, Randy Wootton, CEO of Maxio, speaks with Nicolas Boucher, an expert in AI Applications in Finance. Nicolas shares his rich career trajectory from PwC auditor to a corporate finance leader, and his eventual transition into AI and finance training. He sheds light on the immense potential of AI in streamlining financial processes and enhancing decision-making. Nicolas and Randy discuss automated invoice processing, real-time financial reporting, and leveraging AI for strategic finance operations. They also explore the transformative impact of AI on business efficiency, explaining its practical applications and the importance of continuous learning for finance professionals.
Video transcript
Randy Wootton (00:04):
Hello, everybody. This is Randy Wootton, CEO of Maxio and your host of SaaS Expert Voices, where we bring the experts in and around the SaaS segment to talk about what’s going on today and what’s unfolding for tomorrow. I’m super excited to have Nicolas Boucher, who’s joining us from Stuttgart, Germany to talk about AI and finance.
(00:23):
Nicolas has some incredible background experience, was in audit at PWC for seven years. He worked in the corporate sector for eight years at what would be the French Boeing. We’ll talk a little bit about that, his FP&A experience when he was actually head of finance. And then, he did an interesting shift into corporate finance and became a keynote speaker for the past two years and has done a deep dive into AI, specifically tied to finance, and has started something called the AI and Finance Club, which we will talk a little bit about. So welcome, Nicolas. Thank you for making the time.
Nicolas Boucher (00:55):
Yeah. Thank you, Randy, excited to finally show to people how we can use AI in finance because that’s a topic that is in all of the accountants and corporate finance people and CFOs. And I have the chance to spend my day on this topic, so I’m really happy to share that with you and all of your audience.
Randy Wootton (01:16):
I’m very excited. It is, we’re talking about AI in general, revolutionizing all the functions, and each function having to figure out what are they using off the shelf versus what are they creating on their own and how do they actually think about what they do on a daily basis is very different. It was like night and day, a sea change in terms of how things are going to be done going forward.
(01:39):
But before I go there, I’d just like to go back onto your background a little bit because you have one of these interesting career shifts, backgrounds where you started in audit, that’s a classic start, and then you went into the corporate sector. Can you talk a little bit about that transition from audit into corporate, and what you were hoping to accomplish? And you spent eight years there, so clearly you had a very successful career. What was it that you got out of that sector? And then we’ll talk about your shift into the training sector.
Nicolas Boucher (02:06):
Yeah. So I think, as a lot of us who went through the Big 4 career and audit career, it was kind of the door after the university and the college to continue to learn and we didn’t know where to go, but we are still eager to learn a lot. And that’s what I got from auditing. I could work in Luxembourg and Singapore, many different clients. While it’s already fast, you get a really culture of working hard and working well.
(02:36):
But then, I never wanted to become a partner, and I was interested also to have an impact for companies. And when you shift from audit, where you look at the past, and you go to FP&A, where you look at the future, is a really big shift. And for me, after I did kind of three experiences in audit because I worked in a cycle of two to three years, and so I did three cycles there.
Randy Wootton (03:01):
Right.
Nicolas Boucher (03:01):
I was done with it, that was two experiences, and also family was also a priority. And with my wife, we both decided to shift to more a corporate side where there is more sort of work-life balance, even though I was still working hard.
Randy Wootton (03:20):
All right. Yeah.
Nicolas Boucher (03:22):
And I found a company in Germany because my wife found a job here, which I already did a training before, an internship when I was younger. And so, for me, I was switching I think five parameters when I moved from audit to corporate. First, I changed jobs, so I was no more an auditor, but I became an FP&A manager. Then I moved country, so from Luxembourg to Germany. Then I moved also language, so you have really in Luxembourg we’re speaking English, French, a bit of German. And then, in Germany I was only talking German and I was not yet fluent in German. And then, I also changed industry, because in Luxembourg and Singapore I was working more on finance institutions, so insurances and funds. And finally, I changed culture, because you went from Big 4 to a big defense and aerospace company, which is more long-term and more kind of almost public companies or institutional companies. So that was a lot of shift.
(04:31):
And what I learned is if people are also in the change process, on all of my career, I learned that you have to wait six months. And you should not be like, because you start, you have the honeymoon phase, everything is new, everything is good.
Randy Wootton (04:48):
Right.
Nicolas Boucher (04:48):
And then, after the two, three months you’re like, “It’s really hard.” You don’t have your routine, you don’t have your colleagues. The language is also hard, and you need to wait six months. You need to wait six months to make a point, “Is it really for me or not?”
Randy Wootton (05:03):
Yeah.
Nicolas Boucher (05:04):
Doing a point earlier and being not happy is too early. And at six months, you actually notice a lot of things becomes natural. When you speak, you don’t have to find your words. When you walk across the company, you find the offices, you know your routines, you know what you have to do, you have learned the job. And so, that’s one advice for everybody changing jobs, changing countries, changing language, changing maybe culture, wait six months.
Randy Wootton (05:31):
That’s great. It’s funny, because I’ve thought about in my career there are three variables to think about as you hit those career inflection points is function, industry, and location. And then, you layered in the other two as well as far as language. I just couldn’t even imagine changing all of those at one point. I had an old boss tell me once that when you’re hiring new employees, you need to really wait until 90 days. It’s like a broom. After 90 days, the broom is broken in and you’re going to find out what you really got versus what they said they were going to do in the interview. But I think to your point, if you’re the individual making the change, giving yourself six months to decide, “Was it a good change?”
(06:12):
I know in this job I came out of 25 years in go-to-market, and I’d been CEO twice before but it was always in the go-to-market category and minor shifts between ad tech to market tech to sales tech, service tech, but all on that go-to-market front. And now, moving to the office of the CFO, I was super deliberate about that in terms of wanting to have a new set of challenges, a new set of intellectual stimulation. But to see if the patterns of running a company in go-to-market was similar to running it in the CFO, it’s taken me longer than six months, Nicolas. I’ll tell you that. I’m still learning stuff about how you guys in finance think about the world and how we can sell to you.
(06:54):
And maybe it’s just I’m older and it’s taking me longer to learn, but I do think there is this opportunity when you hit these different stages in your career to do that assessment. Are you happy? Are you being challenged? Are you interested in the problem set? Or do you want to mix it up? And you did that again. So you were FP&A, Head of Finance, and then you decided, “Well, now I’m going to go off and be a corporate finance trainer and a keynote speaker.” What were the ahas and insights you had from your experience either at Thales or at PWC that led you to believe this was your next path?
Nicolas Boucher (07:28):
So actually, it started even before I worked. I noticed when I was younger like a teenager, I was getting a lot of energy from coaching young kids at soccer, football. And then at university, I remember the accounting classes that we had for me, like math and accounting, I had facilities. But I had a lot of friends, it was more difficult for them. So I actually learned thanks to them because I wanted to help them. So I told them the concept and thanks to them I got a good grade because they forced me to learn.
(08:03):
And it continued also in audit when I really loved the part when you become senior and manager and you coach people. And what I noticed is this is something I get a lot of energy from. But in my corporate career, there is only a certain amount of people that you can coach. You have a small team of 10 people, then you can maybe expand to other departments. And also, I didn’t get really a mandate to coach that many people. And there is a technology that exists right now where when you can spend the time to explain to one people, actually you can spend the same time and explain to 1 million people at once.
Randy Wootton (08:44):
Right.
Nicolas Boucher (08:45):
And so, I turned to LinkedIn first because I wanted again, the network of PWC and the network of best practices to learn for myself. And then, when I noticed that I had also something to share, I started to post. I started to see that I had an impact. And there was really one day where I noticed that I had something different is that I was using visuals to explain complex problem. And it resonated to a lot of people. I have the Finance Cheat Sheet, 26 million people saw it.
Randy Wootton (09:22):
Oh, my gosh.
Nicolas Boucher (09:24):
When I read this happened and now I have almost 1 million people on LinkedIn, and across all media is more than 1 million people, so it’s crazy to think about that. But I really thought, “Okay, there is something bigger and something also that gives me more energy than just checking invoices or working on a budget.” And I also noticed a lot of people don’t have the opportunity that you and me got. They were not born in a Western country. They maybe didn’t get a proper job, proper education. They didn’t maybe get a mentor that helped them. And so, I wanted to share that with the maximum amount of people. And for this training, also digital trainings, corporate trainings, was the opportunity to do that.
(10:12):
And working in finance, I was really limited to only finance. But my best role in finance was when I was part of the management and I could do a lot of things that were not finance.
Randy Wootton (10:24):
Right.
Nicolas Boucher (10:24):
And I have a mind like a really entrepreneur spirit, because my dad was an entrepreneur and I think I saw that and I learned so many marketing and entrepreneurship and networking and sales. And for me, it was frustrating not being able to do that so I think going the entrepreneurship journey helped me get a lot of energy from doing many different things.
Randy Wootton (10:50):
Well that’s great. In my own background have been a teacher and a coach, and I can totally appreciate that sense of just fulfillment you get from working with people and helping them to discover something for the first time as well to your earlier point, that in teaching it, you learn it better I think. Right?
Nicolas Boucher (11:08):
Yeah. One friend told me that teaching is learning twice.
Randy Wootton (11:11):
Exactly, teaching is learning twice. Yep, absolutely.
Nicolas Boucher (11:16):
And to come back to the story where what was really the, I would say, the aha moment was two years after my dad passed away, there was one day the energy was really… His birthday, not birthday, but the day of his death, I don’t know how you say that in English. And I don’t know where it came from, but I really got a big force to push the button to say to all of my audience, “I’m starting to monetize all of my trainings and I’m going to build my first course.” And I just sent out, it was maybe 500 people, and I thought, “Okay, if I get 10 people that will buy it, then I will do the course.” And I got 10 people, and then after, it took me two months to do my first course. But it’s crazy sometimes how I don’t believe in this kind of things like this, they are spiritual, but it was really on this day another energy that helped me push the button and say, “Okay, I go for it. And I try it.”
Randy Wootton (12:18):
As my mother would say, “You opened yourself to the universe and the universe spoke to you.” So well done and congratulations on all the success you’ve had. I think this interesting combination of skills and experiences, but also your inclination and your ability to create pictures, right? I think sometimes finance folks, they sit in Excel and Excel speaks to them, but it’s hard for them to translate Excel into concepts that other people can process. I was an English major, and with my CFO, I’m always like, “Look, I don’t want to look at the model. What I want to see is the pictures and the trends and help me see. Because I can see patterns. It’s hard for me to look at numbers and identify the patterns without those images.” And so, I think a picture is worth a thousand words, but the more when you’re doing presentations that you’re able to give visual representation of interesting data.
(13:08):
I think of the map of Napoleon’s march to Russia and how that, do you know what I’m talking about? That one map and it shows the size of the force at the beginning of the march going all the way in and then coming back.
Nicolas Boucher (13:19):
Yeah. Yeah. Really seeing, yeah.
Randy Wootton (13:20):
Yeah. And so, it captured all these different components in it in terms of size of the army, the weather, and then just the attrition of the army through death and starvation. But it was a massive amalgamation of data represented in a really compelling way. And that picture has stayed with us for 200 years or whatever it’s been. So your ability to bring that together is really extraordinary. And so, now you’ve taken-
Nicolas Boucher (13:43):
Actually, there’s really science behind that, especially now that all of us in finance, we don’t have a problem anymore to have enough data. Data is there.
Randy Wootton (13:44):
Right. Right.
Nicolas Boucher (13:53):
We get a lot of data. And so, the problem is we pass through the data to the management and then we don’t help. And there is a science that says that the brain registers and captures an image 60,000 times faster than a word or a figure. So if you really want to… An advice for everybody working in finance and wanting a bit to be noticed by the management, thinking about the Napoleon map, send the graph that the management will remember and they will, when they talk to you, say, “Oh yeah, like this graph, this trend.” This is what will stick, not a long email, not a nice report. The graph will stick.
Randy Wootton (14:39):
Absolutely. My CFO does our all-hands, and one of the things that he does on his own but I reinforce, is draw circles around the numbers that you want to talk about and draw lines showing it’s going up to the right. And then, everyone will see all the different things and get it. “Get it, it’s going in the right direction” or “Uh oh, it’s not going in the right direction.” So that’s great.
(15:01):
And so, then you’re training people on FP&A in particular and how to represent business and participate in conversations at the management level. And then, the AI revolution takes off. What was it? Everyone’s talking about AI today, right? It’s almost passe. But what was it that you saw with this opportunity as a finance professional and being at this intersection of FP&A and training and just your natural inclination to want to go deep dive into this? Because you’ve been in it for several years now, but even before it became popular. So what was it that attracted you to it? And then, we’ll talk about the use cases that you’re seeing that are most applicable in finance today.
Nicolas Boucher (15:40):
Yeah. So all of my career, I really used technology to leverage my work to be faster but also provide more value. And instead of spending hours and hours stuck on a file, trying to automate to be faster and get home early to have fun. So I already saw technology like this and I’m really curious always on where is the technology going and how can we use it for ourself for work? And when on November, 2022, OpenAI released ChatGPT, so the version that everybody knows now, but November, 2022, I think I got the access one day after it was released.
Randy Wootton (16:22):
Wow.
Nicolas Boucher (16:23):
And when I got access to that, you know how it felt, Randy? It felt like when you’re a kid and you get this video game at Christmas from your parents or this Barbie house that you cannot stop playing with this. And so, I got access to this and I was like, “Wow, this thing is crazy. You can do so much things.”
(16:44):
And I felt that somebody was going to take away from me this new toy and that it was too good to be true. So I was just, then trying everything that was coming to my mind and then first for myself, then for business, then also for my content. And I started also to research if somebody else was looking at how to do it to use it for business. Because a lot of people were using it to write a poem to their wife or to write a rap song in the style of Mozart, but nobody was really, I think at the beginning, picking up what does it mean for business?
(17:24):
And when I did my research, I didn’t find anything. So I was thinking, “Okay, if nobody’s going to do it, I’m going to spend a lot of time and try and try and also connect.” Because as an auditor and working in a big corporation in different roles, I saw a lot of different roles in finance but also other departments. And so, I know what in finance we need and where we spend our time. And so, I tried, “Okay, how can you use it to write a procedure? How can you use it to help yourself with your Excel file? How do you use it to create scenarios? How do you use it to help you create a cash action plan? How do you use it if you want to write a code for a program, even if you are not a coder?”
(18:12):
And really quickly, the more I was trained, the more I was documenting that. And in February, 2023, I launched a PDF guide that was sold I think 4,000 times. And now I helped, I think in total, 5,000 people with my courses, with my workshops, with my webinars. I help companies like Mercers Benz, KPMG, training their teams. And my goal really is I believe in finance, we are the champions of Excel, we are the champions of ERP.
Randy Wootton (18:44):
Right.
Nicolas Boucher (18:44):
We are even often really good at PowerPoint and Power BI and some of us VBA and Python. And I do believe that we can be that good and be the champions of business in AI. And my goal is to bring that to people, because we have the culture of using technology for our job. And with AI, there is a lot of data that we can make sense of and I think finance is ready to use AI.
Randy Wootton (19:13):
I think you’re onto something in terms of finance are probably the most fluent in data, business data, right? Because they use the ERP systems and they’re grinding all that out and they’re presenting the reports for the audit, et cetera. And so, then the question is can they take the leap to using a tool that’s more black box, where they have to trust that there isn’t going to be generated scenarios that aren’t rooted in the reality?
(19:42):
I think one of the things when I talk to CFOs is why they like Excel is they can look at the specific cell and say, “I saw that number and I can trace it. Even though I have 50 tabs open, I can go for my summary and drop through down into all the other tabs.” And when I’ve talked to CFOs so far it’s been, “Gosh, I just don’t know if I can trust the output of the model,” so the hallucinations. So how have you addressed that to make people feel comfortable?
Nicolas Boucher (20:10):
At the beginning, first it was each time you were trying to do even the simplest calculations, you could see from your eyes that there was a problem; that the model that we are all using, so the generative AI models, are by definition generating something, so creating, not calculating. So it will give you the most probable output based on your input. And if you start giving a lot of numbers, it’s not an Excel file that you have in front of you. It’s just a model that will try to give you the best answer. And because of this, you cannot rely on the computation because it’s not made to compute.
(20:52):
So once you have understood that, that the model is not there to compute, it’s actually there to help you compute. It’s there because it knows, “Okay, if you want to do a cohort analysis on your SaaS sales, well here are the three or four metas that you could do. You could do maybe a cohort by month. You could do a cohort by type of products or by that type of promotion.” And then, it will explain to you how to run that either in your Excel file, or if you have a lot of data, in Python.
(21:25):
And then, while you do that, you are not giving any confidential data, you are just explaining your problem. If we two meet on a barbecue on Sunday, we are two competitors, but we are friends since a long time and we work as CFO. And I can ask you, “How do you do your cohort analysis, Randy?” And then you can explain me how to do it without giving me any confidential information. And then, I go back after this talk being smarter, and then I will try it on my own data and I will get the value out of it, because now I have a guidance on how to do it. And on top, then it’s like, I talk after to my best friend who is expert in Excel and I’m like, “How can I do that in Excel because I have a really hard time?” And then, ChatGPT or Gemini will do that as well. They will tell you, “Okay, that’s how you do a heat map in Excel, because you have to do a lot of conditional formatting.”
(22:21):
And then, after you talk about your best friend who is really good at storytelling, and then, “Okay, how can I present that to my management because it’s a lot of data?” And then, you will also get a lot of help in this, and you have all of this in the chat bot which has all of these skills. But you can only make sense of it when you ask the good question.
Randy Wootton (22:42):
Right.
Nicolas Boucher (22:42):
And that you also challenge it and you interpret it and you use it in your own environment.
Randy Wootton (22:47):
Great advice. Have you also found… One of the things, we had someone who does this for marketing come in and train our marketing team on how to use AI. And it was a woman, and she said, “Just say please and thank you.”
Nicolas Boucher (22:58):
Yeah. I do that just intuitively because it’s like chatting with a friend or chatting with your team. You always say please or thanks. So it’s just not changing your language because it is just like a chat open and doesn’t really matter. Yeah, adding please doesn’t hurt.
Randy Wootton (23:22):
No, but what she said was it in fact helps, because the AI is trained on language. And the people are saying please and thank you, to your point, in their language in the way they interact with each other. And so, it’s more responsive. I thought it was baloney, but she actually said, “No, no, no. It pays to play.”
Nicolas Boucher (23:41):
Well, thank you is that. Thank you is a permission to the model that don’t swear is really what you expected.
Randy Wootton (23:47):
Yeah.
Nicolas Boucher (23:47):
Rather than choose to save time, rest assured that it will just generate another answer to please you, because the model always want to please you.
Randy Wootton (23:55):
Yeah. And that was the other thing she said, is tell him that it’s really, really important. You’re doing a board presentation. You’ve got to do it right. Give it the extra college try and it’ll come back with yet another answer. And so, I do think there’s this interesting dynamic of it is a machine, but because it’s trained on human relationships, you interact with it in a human way. It’s been fascinating.
Nicolas Boucher (24:17):
A chicken dance. So what you can do if you get an answer, let’s imagine you ask help to get a cash action plan. First because cash action plan is so wide, you can do so many things and you are limited by the context window, so meaning how much output you can get can get from the model, then you will only get a really broad cash action plan. Which if you have experience, you are thinking, “I don’t learn anything here and it’s not going to help me.”
(24:48):
But the technique is to go down two to three levels until you get something super practical. But once you get something that you want to use, then you can ask ChatGPT or Gemini, “Look, review your output, and rate yourself from zero to 10 on how much practical it is and how much specific maybe to a SaaS company it is.” And then, it will give you an answer saying, “Oh, I think it’s at six on 10 or seven on 10. And then, specific to SaaS, maybe on three to 10 because it didn’t know until now that it was a SaaS.” And then, you ask, “Okay, change your answer that you get a 10 on both.” And that is a way to really like 2X or 10X the output you get, because then on these two dimensions, which is practicality and SaaS specifics, it will give you really a much more practical and specific answer.
Randy Wootton (25:48):
Interesting. Well, with that, let’s segue into some of the use cases. We have five, we’ll see how many we can get through. This was based on the pre-brief we talked about and the ones that you thought were going to be most relevant for finance and accounting. Let’s start with automated invoice processing and accounts payable. Help us understand how you think about AI helping in that process to drive efficiency and effectiveness.
Nicolas Boucher (26:14):
Yeah. So now we’re talking about AI and not Generative AI anymore.
Randy Wootton (26:18):
Right.
Nicolas Boucher (26:18):
So in AI you have a lot of different technologies. And one of them is OCR, so optical character recognition, meaning that if you give an image to AI, and that’s something that we have already since a year, so you did have to wait ChatGPT for this because it’s not connected. But you know when on your phone or even when you scan a PDF, it can recognize the text? Well that’s the OCR.
(26:47):
So imagine now a lot of companies and a lot of accounting departments, for years, they have gotten an invoice per mail. Then somebody will have that on their table, and then process that directly in the ERP. Well, the first step is to say, “Okay, I can scan the invoice and let all of the fields already in a pre-formatted table, where after the human will say, ‘Oh, when you see here invoice number, then I map it to my ERP invoice number. When you see invoice amount, I map it to invoice amount.'”
(27:24):
So that was already existing the last 10 years, which is something where you always needed a human interaction for the first invoice. Then after, when the same invoice was coming back, then the model was already trained to understand this invoice, so then it was automated and then processed. Then when you start receiving that by email instead of mail, then you don’t even need somebody to scan it, you process it directly.
(27:53):
But now thanks to NLP, so NLP is natural language processing, meaning what we do with ChatGPT, meaning it understand the words, so you don’t need a human explaining that, “When you see invoice number, you have to map it in the ERP with invoice number,” because then the NLP technology will recognize and see and say, “Okay, at 99.9999%, if I see invoice number, I can map it to invoice number in my ERP. And now thanks to that, you almost don’t need any humans in the loop. And you can do that not only on invoice but on receipts.
(28:30):
So imagine all of these mini receipts that are really hard to read, sometimes also handwritten or in different languages. Because now you have also a lot of transformers that will understand Chinese and if somebody is going to China and going after a late night with their clients and inviting them on, I would say on a ladies bar or something like this, “Well, maybe the eye content will not notice that is a ladies bar because it will just book it.” But the tool will recognize it because then it will see, “Okay, based on the name or based on my research on internet, I see that it’s not in line with the expense policy, so I deny it.” And that’s also how you can process things faster, avoid to spend money on things you should not spend. And also, then the finance team is not there to do mundane tasks but more focused on the problems. And that’s also for everybody. It’s better, because you do work on better tasks than tasks that nobody was interested to do before.
Randy Wootton (29:32):
I think you’re making two really, well, a couple interesting points. One is people use AI, and they mean lots of different things and under the category broadly of machine learning. In my first company back in 2015 I joined, it was part of the first generation of AI where we were using logistic regression analysis building models. But it took us, we had 30 data engineers, data scientists that were building the models, and then we had a set of folks that were working with customers as well. And so, you had this level of abstraction between the user of the model and the model. And so, what’s happened now with the Generative AI is people can interact with it directly. But also, you have these other technologies that have been out there like OCR, which are now becoming more accessible because of NLP and just the interface.
(30:16):
You had mentioned a couple of companies I think for this specific use case that you would recommend or are you familiar with, one was AppZen and the other one was Glean.ai.
Nicolas Boucher (30:25):
Yep.
Randy Wootton (30:25):
Do you want to talk at all about those or any other ones that you would say, “Hey, if you wanted to automate invoice processing and accounts payable, here are a couple of firms take a look at.”
Nicolas Boucher (30:34):
Yeah. So AppZen, for example, was the example I just gave you about this receipt where you go in China, if somebody goes in China. And so, they have their own model where they will read and train the model on your own expenses policies, and then will flag everything or refuse everything that is not part of the expense policies. And they have also been in the market longer than ChatGPT. So they are not like all of these ChatGPT rapper that pretend to be AI, but they just connect themselves to ChatGPT and get the output from ChatGPT.
(31:11):
Glean.ai, the advantage is that they read everything that is on the invoice. So I don’t know if you have experienced that, but when there is a problem with one invoice or when costs increase or when you want to search for cost reductions, then your general ledger is not enough. You have to open the invoice and start to compare invoices over the last two years and start to go into the details. And for this, somebody is either first searching for the invoice, then opening it, then writing all of these details in Excel, and then comparing if we are not over-consuming the service or if we don’t get the increase of price that we didn’t ask, to challenge the invoice.
(31:56):
With Glean.ai, it’s again using the OCR recognition to understand the detail of the invoice and then map that months after months and create an insight, and then flag also the problems and help people to validate invoice because everything is there. They don’t have to go through the 30 pages of the invoice. Already the summary will be on the report from Glean.ai.
(32:18):
And the third one, if you are a small company or medium size and you don’t know which tool to use, but you use something either like Microsoft Azure or Google, they already have those model where you just connect your folder with your PDFs and they have models that for Microsoft Azure is called Document Intelligence. And everything is already made where you just choose the option invoice or choose the option received, and then it will extract directly all of the relevant information that are typically on invoice. Because it’s trained on so many invoices that it will recognize it.
(32:58):
And the advantage is you don’t need an IT team for this. I would say with an appetite for technology, you can do that in two hours.
Randy Wootton (33:07):
Wow.
Nicolas Boucher (33:07):
Even if you not a tech person. And that really removes the barrier of entry for a lot of people to go from, “I have no process of digitalizing my invoice” to one afternoon after, “Okay, I start to have all of my invoices in a flat table.” Then the next step is, “How can I connect that to my ERP?” And then, if you have an ERP with a easy API, then you just transfer this flat table with the API and it books for you all of the expenses in the ERP or in the accounting system.
Randy Wootton (33:42):
Well, maybe that’s a good lead into one of the other use cases, and we’ll probably just get through one more, is real-time financial reporting and analysis. So if you start out with processing the invoices in the accounts payable and you’re able to automate that process and then have it go to the ERP, what are you seeing in terms of AI that are generating real-time financial reports offering up-to-date insights into revenue expenses and profitability?
Nicolas Boucher (34:07):
You know how until now you have your revenue details in another other ledger and FP&A has to make sense of it, make reconciliation? And then, you need two or three days to understand what are the sales you did in the details and why did you have variances? And now you have Puzzle.io, which is done by Sasha Orloff, who started with the premises that he’s not going to change companies’ sublegers, because either your subledger are in the right tools or not. And if they are not, then he just says, “Okay, I’m not going to work with you.” So he works with startups and what he does, he says, “Okay, we want four conditions. The first, that you are in a bank where we have the API, so Mercury or I think there are a group of banks that you can have easy APIs. So like this, the cash rack is super easy and everything is uploaded and in real time inside your accounting system.
(35:14):
The second one is on payroll. So with Gusto, they also have API, and you get real-time information on your biggest expense, which is salaries, because 80% of our expenses for most of the companies are headcounts and salaries. And then, on payables they have a connection with rent, and then the last one on revenue, connection with Stripe. So with all of these four subledgers, then they can create real-time accounting and all of the SaaS KPIs that everybody needs and are calculating on a Excel file. They do that straightaway when they integrate the four subledgers. And on top of that, they use AI for categorizations to help also generate insights.
(36:06):
But it’s mostly first the value in the integrations. And AI is not solving everything. It’s first do you have your data? And then, AI can help in the mapping, can help in flagging errors. But it’s mostly like do you have the data? Do you have the right data? Then can you map them through integrations where nobody is touching anything, it is just coded once, and coded the right way that you can get value from it.
Randy Wootton (36:32):
Well, I was going to say, I’d be remiss if I just didn’t do the Maxio advertisement here, because Maxio is a revenue recognition, revenue management system. And so, to your point, people that are trying to create the finance tech stack of the future where they do connect via APIs, but then sitting between the CRM and the general ledger to pull all that data together to do the revenue recognition. And then, for SaaS companies in particular, to provide the output in terms of the operating metrics, your gross retention, your net retention, your MRR roll forward, all those different components. So I’ve got to talk to this guy at Puzzle.ai. We’ll cover off on that separately.
Nicolas Boucher (37:07):
Yeah.
Randy Wootton (37:07):
So we’re kind of bumping up against time and I wanted to get through these other ones. So maybe we’ll do a V2 of this, Nicolas if you have time.
Nicolas Boucher (37:08):
Sure.
Randy Wootton (37:14):
But just to close it out for our audience, I’d love to talk about the speed round. So the speed round, there are three questions. What’s your favorite metric and why? What’s your favorite book? It could be business or personal, and why. And then your favorite influencer, someone you’re following or you think either in the area of AI or in the area of finance that you think is really writing original content versus just putting a spin on other stuff. So favorite metric, what’s your favorite metric?
Nicolas Boucher (37:42):
Well, I like to understand in a company what is the salary by headcount?
Randy Wootton (37:48):
Ah, okay.
Nicolas Boucher (37:49):
Sorry, the revenue by headcount.
Randy Wootton (37:49):
Okay.
Nicolas Boucher (37:49):
Revenue by headcount, because it helps me understand if the company has a good margin or not, because usually either… And not a lot of companies are trading companies, so if we forget that, a lot of companies, the value is created by the people. And so, revenue by headcount I think is a good indicator to see if the company is profitable or not. Because usually people will tell you the revenue, they will, and headcounts you can find out everywhere. But you will never know about debits, so it’s a good way to feel if the company is profitable or not.
(38:23):
And then, what I also like is for people themself. In my team, often I was explaining to them, “Look at how much value you need to generate. If you look at your salary and the revenue, if you get a salary of 100 and you have to think about all of the other costs, then it means that you need to generate 150 or 200 as value.” So to make people stick that to themselves and feel that they are part of the equation. Because if not, everything is numbers for people that they cannot touch it. But if you tell them, “Okay, if you were a loan, you have to create $150,000 or $200,000 value to cover your salary.”
Randy Wootton (39:09):
That’s great. I know that in my previous incarnation as CEOs, we’ve always had a focus on ARR per headcount or ARR per FTE, so almost like a proxy for revenue. It’s a little bit different. But working for a PE company, they have been laser focused on cost per FTE, and how do you think about creating more efficiencies that drives the EBITDA? Because to your point, such a large percentage of the costs are salaries. And so, how do you think about that? And I think in this hybrid world now where people can work anywhere, it does create opportunities. And AI also if you’re adding augmentation, I know I’ve seen slides going around amongst CEOs I talked to that invest in firms, VCs and PEs are expecting 30% efficiency starting next year driven by AI augmentation. And so, this idea of how much leverage are you getting from AI is going to be critically important for people’s success going forward. Okay. So that was the metric. Favorite book?
Nicolas Boucher (40:11):
So I have to go something a bit non-business and non-finance.
Randy Wootton (40:11):
Okay.
Nicolas Boucher (40:18):
But still help us understand the world is Sapiens.
Randy Wootton (40:21):
Oh, sure. Right, yeah. Great. Yeah.
Nicolas Boucher (40:24):
Yeah. Because I think when I read it, first, it’s fun to read even if it’s big, but you can stop at any time and come back. And I think it teaches us so much about why are we what we are today and what happened before.
Randy Wootton (40:36):
Yeah.
Nicolas Boucher (40:36):
And it’s a great way to step back and make sense of everything that is happening right now.
Randy Wootton (40:44):
Awesome.
Nicolas Boucher (40:45):
Or personally as a person.
Randy Wootton (40:47):
Did you read the follow on? It’s by Noah Harari I think.
Nicolas Boucher (40:47):
Yeah. Yeah.
Randy Wootton (40:47):
And then-
Nicolas Boucher (40:47):
I think it’s Homo Deus or something like this?
Randy Wootton (40:59):
Yeah, something like that. Yeah, yeah, yeah, Homo Deus. You’re exactly right. I didn’t read that one, but Sapiens was one of those ones, I can’t remember when it came out, but it’s absolutely transformational.
Nicolas Boucher (41:06):
No, I didn’t like the Homo one. Yeah. I didn’t like the Homo Deus because I think it was going on something where it was not that impacting. But yeah, the first one was so good.
Randy Wootton (41:19):
Okay. Sapiens, yeah.
Nicolas Boucher (41:21):
I had also high expectations.
Randy Wootton (41:22):
Yeah. And so, I just looked it up quickly. It was published in Israel in 2011, and then came out in 2014, so it’s been 10 years, which is hard to believe. But yeah, I’ve got it down. This a good one maybe to pull that back up for a summer read, because to your point, it was really just eyeopening when I read it. Okay. Favorite influencer, so someone that you like to read in the morning. You wake up in the morning, you grab your cup of coffee, who are you reading that’s providing interesting ideas and insights?
Nicolas Boucher (41:51):
So it’s less reading and more watching.
Randy Wootton (41:54):
Okay. Watching, than.
Nicolas Boucher (41:56):
Two person, a bit in the same area, is Scott Galloway.
Randy Wootton (42:00):
Of course. Great, awesome.
Nicolas Boucher (42:02):
Because he has a really spiky point of view, and you might accept or not what he says, but I really like that he has his own point of view, especially on two things. Basically, technology is not good for everything and that we all let TikTok coming into our words, but TikTok is managed by Chinese.
Randy Wootton (42:26):
Yeah.
Nicolas Boucher (42:27):
And also, I think it’s more American-focused, but on the youth. The second one is, I guess not a lot of people know him, is Shawn Kanungo.
Randy Wootton (42:36):
How do you spell the last name?
Nicolas Boucher (42:38):
Yeah, so K-A-N-U-N-G-O. And I talk to him. So it’s a guy who is doing a lot of keynotes, and he has a really spectacular style of doing keynotes on innovations.
Randy Wootton (42:54):
Okay.
Nicolas Boucher (42:55):
And you need to look. His keynotes are, I think, the best in terms of experience. And I love the fact that he makes, I would say, both key results with these keynotes. First learning, so you learn something and you reflect. But second is also really entertaining. And for me, who is also doing keynotes, is I see that as an example and a path to follow; that if you take time of people, if you are in front of people digitally or live, and even now in this podcast, if you take one hour of people or one hour of 1,000 people, then you owe to the people that they have fun and that they learn something. And for this, I think you have to work for that and you have to be conscious about this. And that’s why I really like those two, Scott Galloway and Shawn Kanungo, because they have each in their own style great learnings for me, but also for everybody who is just watching them. They are really entertaining and insightful.
Randy Wootton (44:06):
Wow. I could not think of a better way to wrap this episode because that’s our ambition. I don’t know if we will achieve it, but it would’ve not been because of you, because you did a great job in sharing some insights in terms of your background and experience and what you’ve seen unfolding with AI and its specific use cases. So thank you, Nicolas, for your time. It’s really been a great pleasure.
Nicolas Boucher (44:25):
Thank you, Randy. And call to action to all of the audience.
Randy Wootton (44:28):
Oh, that’s right. That’s right.
Nicolas Boucher (44:31):
If you learned something today, share it around. Because I think if you are listening to that, it means already you are in the top 1%, and you owe to the others to bring them with us to the learning path. Because AI is going super fast, and there are some people that are maybe either not as lucky or maybe don’t see that, but you need to bring them with us and teach them and show them the value of it and also the risk and limitations.
Randy Wootton (44:59):
Right. And so then, sorry about that, the call to action. People can find you on LinkedIn, they can join the AI and Finance Club, where if they wanted to learn more about what’s happening in AI at the intersection of finance. You have a ton of materials there they can access and they can be part of the conversation. And to your point, they can pay it forward and help all of us get better together.
Nicolas Boucher (45:20):
Exactly. Yeah. I just wanted, if people don’t want to join us, it’s not my intention for this podcast. I really want to teach. But in the AI Finance Club, it’s really the place where if you are a CFO, if you are a professional CFO, if you are just touch anything about finance and want to learn, but you feel that alone is not enough because it takes a lot of time, then you can use me, you can use the experts that are with me. Because we do the work, and then you just get every week just something you have to consume in five to 50 minutes. And one time per month, we also meet in a masterclass where we can learn from each other and I bring experts. And I found that was the best way to learn because we all come from different backgrounds. A lot of people also bring their insight, but all of the questions that we receive, if we answer them, then everybody benefit from it. And it’s great to learn together.
Randy Wootton (46:17):
All right. Well then, that’s where we’ll wrap. Thank you, Nicolas, for your time. I really appreciate it.
Nicolas Boucher (46:21):
Thank you.