Sylvestre Dupont Podcast Transcript
Sylvestre Dupont joins host Brian Thomas on The Digital Executive Podcast.
Brian Thomas: Welcome to Coruzant Technologies, home of the Digital Executive podcast.
Welcome to the Digital Executive. Today’s guest is Sylvestre DuPont. Sylvestre DuPont is the co-founder and CEO of Parseur, an AI powered platform that automates data extraction from documents before launching Parseur. He spent over a decade in consulting where he led complex international projects across diverse industries with a strong background in computer science and extensive global business experience.
Sylvestre combines technical expertise with a visionary approach to simplifying business processes. He is passionate about empowering businesses to focus on what truly matters by automating tedious tasks so they can deliver greater value to their customers.
Well, good afternoon Sylvestre. Welcome to the show.
Sylvestre Dupont: Hello. Good morning to you.
Brian Thomas: It’s great to have you hailing out of the great island of Mauritius. I think that’s amazing. The first time I’m doing a podcast there, so I really appreciate it. I guess this is 53 countries I. If you want to call it that. So thank you again.
And Sylvestre, let’s jump into your first question here. You transitioned from leading consulting projects to founding Parseur. What inspired you to make this shift, and how has your consulting experiences influenced your approach to running Parseur?
Sylvestre Dupont: I. If I’m being totally honest here, it’s a girl that figured that it all, I was dating a girl back in the days in Paris, in France, and things were going well, and she had an amazing opportunity for a new job, but that job was hundreds of kilometers of miles away from uh, Paris.
So that was a bit of a pickle. Then at the same time, my job in consulting was about basically three things. It’s meeting emails and Excel spreadsheets. So it’s fun, I love the job, but something was missing and uh, the creation part was missing. So at the time on the side, was trying to go back to my formal love, which was programming, coding, and I really enjoyed creating things, products.
So I decided to use this excuse of, uh, having my girl relocating at the time, uh, to someplace else to try something new and launch for in business. At the time at Accenture, I saw so many inefficiencies, let’s say, of people having to copy and paste data from documents to some other software or some of the places to make use of that data that have, maybe there is something that that could be improved.
And this is how Parseur as a arose from there. And the idea was just let’s get all the data trapped in the documents and automate the extraction and then push the data where it belongs. So this is, uh, at the same time I had a, now a co-founder, but the, at the time, a very good friend as well that was in the same mindset.
He wanted to try something new. He wanted to try her own business. So we decided to start it together and it took a while, but in 2016, Paso was born back to today. That girl from the time now my wife.
Brian Thomas: That’s awesome. I love the stories. Uh, we typically kick off the podcast with a great story like that, and you found the two loves of your life, right?
Developing, uh, coding. And then of course you found the love of your life, which is now your wife, which is awesome. I I really appreciate that. Sylvester, can you explain how Parseur leverages artificial intelligence to automate data extraction and what advantages this offers over traditional methods?
Sylvestre Dupont: So first and foremost, Parseur was born in 2016.
So at the time there was not much ai, to be honest. When Parseur started, we started it the traditional way. So people would send some documents to us and then we would ask through a visual interface, oh, what data you want from that sample document, and it would grow over it. And we said, okay, understood.
You want this data there, that data here, and we would do it. Great. And, and still today it’s still a, a very major feature of our tool that allows to today reliably extract data visually. But then AI came a few years ago, as everybody know, and that tremendously changed what we could do. There are many three things that we could not do easily before that now AI will allowed us to do.
The first one is that. The AI understands what a document is about. They understand the context, so you don’t have to visually ask somebody, oh, if you receive a, an invoice, and the invoice number is exactly at that spot in the document, but if you receive that invoice, it’s not there. It’s a bit further below.
And then before the people used to have to create what we call templates and, and show the differences for each types of documents. Now, you don’t have to do that anymore with ai. You just say, oh, I want an invoice number. And then you send the full version of the invoice, a aspects or picture. And the AI is smart enough nowadays to understand that.
So it, it’s a massive game of time for our customers. They don’t have to spend minutes or hours even for the legal one to, to configure their parcel county. Just say, oh, I want this data, that data, and off you go. And there was, we were able to go beyond that even. And before we were just extracting data, so it was just, you see an invoice number, that’s gonna be an invoice number, but now is ai because they understand document.
We can go further than that and you can do translations, for instance, say you receive some leads for, with a message from, uh, your broker in the middle of the, of, uh, Miami and you receive some requests in Spanish because there are many, many Spanish speaker there, but you happen not to speak Spanish. You can just ask the AI to translate the message they receive in English if it’s not, and that works very well.
You can also ask the AI to summarize the document in a few sentences. If you, you have some very large contracts and you want to, uh, to, to have a quick overview of it. So really it, uh, what was not even thinkable a few years ago, it’s now possible with AI and in just in a few clicks for, for the users, so that, that’s great.
Brian Thomas: Thank you. I appreciate that. And I really thought your tool before the, you know, really the introduction of AI was amazing with that reliably extractive data, doing that visually, but with ai, as you mentioned, things, uh, it’s just like everything’s on steroids now. Your platform has really taken off. And of course, you know, those complex things like various languages, similar summarizing large documents and, and not having to create templates just.
With that AI is able to parse that information to a higher level, so I appreciate that. Sylvestre, how does Parseur ensure its platform remains user-friendly for clients across diverse industries, and what feedback mechanisms are in place to drive continuous improvement?
Sylvestre Dupont: Well, there is some things that we believe strongly in at Parseur is that we hate complexity.
We are striving for simplifications for every single thing, but that applies to our, uh, user interface that applies to our processes. That applies how we are organized internally. Every time we see too much complexity, we try to remove it or we try not to have it in the first place. So for, for the. Keep our application user friendly.
We strive to remove as post as many settings, Dropbox pages, configuration items as possible from the user journey. And, and of course we hope that, uh, s makes it work because otherwise we, we fail to, to do what we are the customers are paying. As for there is always a, a bell curve. So you start having a very simple app.
You just launched, you have one feature, everything works well. But then. Customers comes in and start to ask for more features, and that’s great. It means they, they, they want to use your tool. They, they want to expand it. And I think the, the game here is to try to see how you can increase the surface area of your app without increasing the complexity of using it.
Because at some point you, uh, if you don’t do it, you end up in a, in an app that can do everything, but is a bit like a gas factory and just people don’t have time for gas factories. They just want something simple that, that does the job. I think this, as I said this, this simplification mindset is at the heart of everything we do at Parsa.
We try to, we are bootstrapped, we have no investors. We keep our team low. We try to automate as much as possible and try to keep as lean as possible.
Brian Thomas: I. That’s awesome. I really love that. I respect a lot for, uh, entrepreneurs that bootstrap because that is one of the toughest ways and sometimes some of the slowest and grinding ways to build a company.
But I really love your focus on customer, the customer experience, and the customer. The goal is to remove that complexity and, and make it as user friendly as possible and, and glad that is your focus. So. Thank you again. And Sylvestre. In your view, how will AI continue to transform business processes over the coming years and what role do you envision parkour playing in this evolution?
Sylvestre Dupont: I. I think, uh, AI will increase productivity tenfold by companies. And the analogy I like to make usually is that for me, information or data is like money. So if you have $100 bill and you put it under your mattress, it’s great. You have $100. It’s good for days maybe, but at the end of the day, it doesn’t serve any purpose.
It, it’s not valuable per se. But if, uh, somebody pays you a hundred dollars for something you did for them customers. Then you can use that a hundred dollars to pay your employees, and then that employee can use that a hundred dollars to, uh, go for a dinner, a nice dinner at night, and then the, the restaurant owner can use that a hundred dollars to buy the products from the distributors.
And the distributors face that money to buy from the farmers, et cetera, and et cetera. And then the same a hundred dollars becomes so much more valuable. Data is exactly the same. You have some data stuck, trapped in your document, on your desk, in your mailbox, on your servers, and it’s great You have data, but if you don’t do anything with them, it’s, it’s useless When the data starts to gain value, when you start to use it and move it from A to B to where it belongs, use it for dering all the data to make business decisions to to, to grow a company and.
AI will increase that flow of data. Like the money well spent is increasing the, the, the, well, the value of the money and AI is going to do maybe a few orders of magnitude more in the data flow, the speed of the data flow in the coming years. And, and that’s a tno shift. I think that’s as big of a shift as the internet or the electricity before, or the combustion engine even before that.
It’s really gonna change a lot of things and look a tremendous amount of wealth, not only for for companies, but for just the, the day-to-day life. And just, uh, for instance, for at Parseur, we, in the last two years, we have processed 10 times more documents than the year before that. So in two years, we have increased of volume flow by 10 x.
And this is greatly because of data, because it, it’s easier now. And even if I see that in the future, hopefully document data, tracking document will be less and less and, and there are gonna be more system interconnected into each other. So there is gonna be less use of parcel, parcel will still be around because there will still be the, the some of, of that data information truck curtain documents and we hope for password to be the main bridge to connect the remaining data in those documents to the rest of, uh, the data work.
Brian Thomas: Thank you for sharing the analogy as well. You know, you talked about AI is gonna increase productivity tenfold, and that analogy used as the value of making that a hundred dollars and how you spend it and the values passed on. AI, as you said, will increase productivity efficiency of a much higher magnitude.
And, uh, we can see that obviously through the work you’ve done with Parseur and it’s grown, uh, significantly over the last couple of years. So I really appreciate that and Sylvestre, it was such a pleasure having you on today and I look forward to speaking with you real soon.
Sylvestre Dupont: Yeah, thank you very much, Brian.
I, I love the invitation and, uh, it was was nice speaking with you today.
Brian Thomas: Bye for now.
Sylvestre Dupont Podcast Transcript. Listen to the audio on the guest’s Podcast Page.
