How AODocs CEO Stephan Donze grew to $60M revenue with a 95 person team in 2026.

Nathan Latka (00:00) Hey folks, my guest today is Stephane Donze. He's the founder and CEO of AO Docs, a French engineer with a master's from a great university. He previously served as VP of engineering at Exelede, which was acquired by another company in 2010. He's got 25 plus years of experience in enterprise software and NLP. Stephane, you ready to take us to the top? Stéphan Donzé (00:19) Yes, nice to you. Thanks for having me today. Nathan Latka (00:20) I, it's good having you and I don't want to bury the lead here. My team told me that you have bootstrapped the company to over 55 million bucks of revenue. Is that accurate? Stéphan Donzé (00:32) Yes, ⁓ we actually ⁓ grouped together product ideas like EODox and a service company. ⁓ Our parent company named Revivor was one of the first cloud integrators in Europe. And we used this ⁓ service side of the business to fund the beginning of EODox until EODox itself became profitable about four years ago. Nathan Latka (00:59) Are you comfortable sharing what just just AO docs by itself does today in revenue and then we'll go get the launch story and the product story Stéphan Donzé (01:05) Yeah, our DAX is about 30 million today. Nathan Latka (01:08) Okay, and what's the corresponding website to the service company or is that embedded up here? Stéphan Donzé (01:12) No, it's not on the website. It's ⁓ REVEVUL.com the E-VOL or E-V-E-VOL. Nathan Latka (01:18) Like that. R-E-E-V-O-L. Stéphan Donzé (01:22) No, no, R-E V-E V-O-L It's a terrible name, you know. ⁓ Nathan Latka (01:30) Okay, all right. So from cloud to AI. Okay, so when did you launch this service company? What year? Stéphan Donzé (01:35) This was founded in 2007. It was the very first partner of Google in Europe when Google launched their enterprise B2B edition. then Eodox was created in 2012, first under Revol. And now we kind of it around because Eodox became much bigger than Revol. Nathan Latka (02:00) Interesting. So before AODocs was launched, what was the maximum, like the biggest revenue year you did at the services company and what kinds of services were you doing? Give me an example. Stéphan Donzé (02:10) We were doing cloud migration. So the companies wanted to move their email files and calendar to the cloud. So it started very small. In 2010, it was doing a couple million dollars of revenue and the two companies grew together. the miracle of our model is that the services activity grew fast enough to sustain the software development activity. And until 2018-19 where the EODOX itself was almost profitable so we were able to launch ourselves. But from 2012 to 2020 there was a synergy between the two businesses so that the profit of the service activity would be growing fast enough to sustain the growth of EODOX itself. Nathan Latka (02:58) And then you became AO docs became independently profitable in 2022. Correct. Okay. Stéphan Donzé (03:03) Yes, correct. in 2020, 2021, we were like half a million in the negative. So we were almost profitable, but we became really profitable, really positive in 2022. Nathan Latka (03:14) You were burning at AO docs $500,000 per month. That keep you up at night? Did that make you nervous? Stéphan Donzé (03:19) The most we did actually was 200,000 a month. we were never really too much in the red, but that was very close to the whole profit of the service activity. So it was tight and we had to become efficient. But the good aspect of this, but at least it forced us since the beginning to be efficient in... R &D, in marketing, in all aspects. We never went into a spending frenzy like you can do when you have a big fundraising. The downside of it, of course, is that in terms of communication opportunities, we never got to make a lot of PR around the big ⁓ fundraising. Obviously, we were a bit low-key for these years, but we focused on large customers and doing ⁓ reliable job and good quality. reputation ⁓ helped us grow fast without even without VC-funded marketing dollars. Nathan Latka (04:24) want to talk more about your growth, but first that update you gave on 55 million was middle of last year. imagine you have an update today. What do think you'll do combined revenue in 2026? Stéphan Donzé (04:34) In 2026, can, I think we can count on hopefully 15 plus percent growth. So we'll see. We have a few big deals that should land at the end of the year. it's hard to say, to tell at plus or minus 5 million. Nathan Latka (04:53) So maybe break 60 million of revenue this year or something like that. Stéphan Donzé (04:57) Yeah, I think so. mean, it's a balance between on the side we have a reselling activity that is going down. in the whole numbers, it might not move a lot, but in terms of generated margin, we expect a big increase. So the joy of that counting. Nathan Latka (05:15) And what are you and the breakdown and let's say you do 60 million this year, is it 50-50 between AO docs and the services company or no? Stéphan Donzé (05:23) Eodox is a bit more than half. Then we have another software company called Talarian, ⁓ which is the maker of ⁓ GPT for Sheet. It's an add-on in Excel to bring AI in Excel and Google Sheet. And then we have the service activity. So the other half is partly Talarian, the other software company, and partly the services activity. Nathan Latka (05:44) Help me out again, Stefan, with naming. What's that third company? How do you spell it? Stéphan Donzé (05:48) t-a-l-a-r-i-a-m-s.io ⁓ Nathan Latka (05:53) Okay, when did you come up with this idea? Stéphan Donzé (05:55) Both companies have similar origins. Through the services activity, we do projects with customers and we realize needs and things that are requirements that are not well covered by the market. So just like when EODOC started, because there was no document management solution available in the cloud, Talion started because there was no... There was need for several functions, sending emails from a spreadsheet, ⁓ merging documents, creating documents from a spreadsheet, then later integrating spreadsheets with AI correctly. So we heard those needs from our customers and we built those apps based ⁓ on what we saw on the market. the services company not only gave us funding for the software, but it gave us a good point of observation to be on the ground. connected to the business users and seeing live what were their frustrations, were their needs that were not adequately covered by the market. Nathan Latka (06:56) I have a confession, are you ready for something? I had no idea you were behind yet another mail merge. I love this tool so much. Look at this email from, I sent you this email in 2016, 10 years ago. Look at this email. Stéphan Donzé (07:00) Yep. You know Yama? There you go. And yes, have Roman Zanari in the loop. Nathan Latka (07:16) I wanted to buy you so bad. It was so simple and so powerful. And when I couldn't get an answer, I'm like, they must just be printing money. They don't need an acquisition. And sure enough, you are probably printing money. Stéphan Donzé (07:20) Yeah, it's still... we are printing even more money. In 2016 it was probably around a million dollars revenue, now we are like six times that. Nathan Latka (07:36) I had offered three to four million back then and I never got a reply. But so has Romain your partner? Stéphan Donzé (07:41) Romain is the founder but now it's another person who's managing the Talarian. Talarian grew beyond just... Nathan Latka (07:48) Well, so did you buy it then? Is this like your holding company? Stéphan Donzé (07:52) No, it's a continuum. was called Awesome Gaps at the time and now we renamed it to Tallinn, but it's the same company than in 2016 when you talked to Romain. Nathan Latka (08:04) I don't understand. did you buy it from Romain or are you a co-founder at Yam? Did you help build Yam? Stéphan Donzé (08:09) No, Wal-Mart was working for us and the company still exists. He went to do something else, but he started the company and the company is still there and it's the same company. Nathan Latka (08:21) So I see. So Yam is a child company under AO docs or your services brand. I see. And you had a, and you appoint to, you appoint a CEO per each of these units. Stéphan Donzé (08:25) It's a brand. It's a brand of Talarian. You have five of all. No, there's a CEO for Talarian. And then under Talarian, the two big things are GPT for Work and IAN. And awesome table for publisher are kind of smaller. But there's a single CEO for all of Talarian, the four products that you see here on the page. And then each product has its own development team, obviously, but there's a single company for the four. Nathan Latka (08:54) Oh, very, very cool. This is so cool. Okay, you are, I wasn't expecting to go down this route with this interview, because I saw your background as engineering, but I have to give you a prop. I mean, you are very, very good at distribution because these apps, I remember back in the day, Yam had like millions of usage or downloads or reviews or something. Stéphan Donzé (09:10) It's still in the top three, I think, of the marketplace for Google Sheet. Nathan Latka (09:17) I'm trying to, let me see if I can find it really quick, hold on. 18,209 reviews, yeah, this is it. 17 million downloads. are you driving growth and distribution? I mean, I don't wanna ask the question about this one, because it's old. How are you driving growth and distribution for this agent? Stéphan Donzé (09:29) It's same strategy. were first in bringing AI in Excel and Sheet way before Copilot, way before Gemini and Sheet. And ⁓ word of mouth, some YouTube marketing, a lot of LinkedIn, but frankly, a lot of word of mouth. Nathan Latka (09:50) Interesting, interesting. Wow. Okay. Okay. Interesting. Stéphan Donzé (09:54) And it's funny because the two companies have vastly different go-to-market. On EODOX, we have relatively small numbers of very big clients, the Arbus, the Air Liquide, the Veolia, and like. While on Tararian, they have tens of thousands of tiny clients. Nathan Latka (10:09) Interesting. And do you have a motion that connects them all? Do you pick off enterprise, you know, employees paying five bucks a month here and try to upsell them to a hundred K plans on AO docs? Stéphan Donzé (10:17) We have a few joint customers, but usually it's different population. The people using Yam, for example, are not ⁓ working in marketing departments while EODOCs, we target engineering, we target IT, we target HR. ⁓ it's anecdotal, but we don't have a strategy except of cross-fertilization, but no technical integration on joint use cases. Nathan Latka (10:46) Okay. Okay. So to sum all that up before we get more of your backstory, 60 million target revenue this year, 30 million of that will be AO docs. Another like, you know, 40 % will be a Talarian and then the rest will be like maybe five, 10 million will be the, the, the services company. Stéphan Donzé (11:00) Services is more like 10 plus, yes. We resell some licenses, so part of the 55 is reselling, then we have integration services, then we have Talarian, then we have EODOCs. So it's a stack of different things. Nathan Latka (11:04) 10 million plus. Okay, interesting. Interesting. So I don't even know where to go with this question. I mean, there's a couple things I'm wondering. First off, you must have incredible focus and discipline because I'm sure you're getting hounded by private equity firms to buy the company and you've resisted for 10 plus years. Stéphan Donzé (11:27) A couple times a week, yeah. I mean, we worked hard to build a profitable company. And now we're at the point where the profit, the habit of the company can be allocated in investment, hiring more, growing, so on, a little bit in returning to shareholders, a little bit of internal things. We do like a... yearly company seminar in Italy. So, I mean, it's fun to be growing a profitable company. don't... And the point is the private equity playbook, which is going to be to do cost cutting, to maximize the profit or to... It's something we can apply ourselves. I why would I sell the company to someone who can do and then reap the benefits of our own work? If someone comes in with a platform that can... distribute the product and accelerate our growth. Sure, why not? That's a different story. But the pure private equity play, which is, take a business that runs well and make it run better. No, I mean, we can do this ourselves. We love our company. We love our independence. We have a great market positioning. Both companies, Talarian and EODOCs, are on the front seat of the AI wave. So we want to take advantage of that. Eodox makes AI possible on critical documents because we have the right platform, the right foundation on which you can put AI agents on reliable information. Talarian is bringing AI in productivity tools to help companies accelerate. So in both cases, we benefit from the AI wave and we want to take it. We don't want to sell the company now that we have a fantastic tailwind. We don't need money, we have money. We have a good pile of cash at the bank. We have profitable business units, so it's good. Nathan Latka (13:31) Can I ask you what was profit margin last year? Stéphan Donzé (13:34) Profit margin on EODARX is 10 to 15%. On the whole thing, 10-ish percent. Nathan Latka (13:37) The whole thing, the whole company. And before we move away from the private equity thing, let me just play devil's advocate for a second, right? You're doing 50, 60 million of revenue. You're very profitable. You you remind me actually a lot from a DNA perspective of this comp. Let me just share real quick. You maybe you've heard of them. but air slate is really interesting, right? They grew by building a lot of sort of small tools that went viral and now they're buying other tools. So the being a devil's advocate to the statement you just made, if someone like this came in, Stéphan Donzé (14:04) Yeah. Nathan Latka (14:08) and offered you, you know, 15 X all cash, your current revenue. So something between 500 million and 700 million to sell the whole thing out, right? You, wouldn't take that deal. Stéphan Donzé (14:17) I don't know, we need to talk, right? Obviously, if someone knocks on the door and then we've had that in the past, we discuss, we talk about what makes sense and so on, but we're looking to say our, there's no exit strategy, our goal is to grow the company. That's why I was saying to me, as the, and I think my colleagues at Talarian and Aurelion are all the same, if someone shows up first, I would ask them, how are you going to grow my company? I don't want to sell it, but I want to grow it. Are you helping me grow? Okay, let's do a partnership. Or you want to acquire me? How is it going to help me grow? Okay, that story I can listen to. But not, I I'm just not just looking at the cash course. If you offer me 10 billion in cash, okay, I will think about it. But it's, I I like my company. I don't want to see it acquire to wither and die. And there so many acquisitions that... failed that I don't want to see that but if you have a platform that helps me grow, yeah, we can talk Nathan Latka (15:19) And Stefan, just to be clear, there's a lot going on here. You're like when we do the consolidated financials, consolidated cap table at the top, you're majority owner. Stéphan Donzé (15:27) Personally not, but I'm part of the owners. Nathan Latka (15:30) Okay, how many co-founders do you have? Stéphan Donzé (15:33) We are the physical people who are manager slash owner, like four or five people. And these are all business partners that have worked with me since 20 plus years. Nathan Latka (15:40) Okay, okay. Wow. Amazing. Okay. Let's go on AI. Everyone's building products in the age of AI. They're wondering when Claude releases the next extension, is it going to kill us? So my blunt question to you with all due respect, why would people use GPT for work when I can just put Claude in Excel? Stéphan Donzé (15:57) Well, it's different use cases. GPT for work is extremely good at ⁓ bulk application. If you have, if you try Claude and Gemini and the likes on more than 100 lines, it's going to slow down and fail. what we're doing here in terms of plumbing, in terms of data management is hard. Understanding the full spreadsheet and running for long time. managing hours and performance. If you apply this on tens of thousands of lines, as it is the case in e-commerce and use cases with very large spreadsheet, it is a tough problem. think Claude has done a good job in terms of UI as integrating for simple use case, but if you have something that's large scale, you need more than that. And we've shown the difference. If you look on YouTube, there was lots of comparison videos. You'll see that that's We measure the difference and there's a stark difference. Nathan Latka (16:54) And what are you doing while, while I'm loading up the YouTube stuff? What do you, mean, I have here, block processing with AI. What are you doing under the hood? looks like you're sitting or doing tool calling this screenshots, Opus 4.6, but what have you built in your tech stack that allows you to manage millions of rows when, when Claude in Excel can't. Stéphan Donzé (17:09) Yeah, we have our own back end that focuses on performance. And if you want to talk about that product, should better talk to Stan, who's the boss of Talion. He'll more to talk about your talks. That's the guy on the picture, that's Stan. Nathan Latka (17:18) Okay, okay, fair enough. Tell me more about the growth. I mean, are you guys actively recruiting community folks on YouTube to do comparison videos? Is there an affiliate program? Stéphan Donzé (17:33) We don't have a format affiliate program, but I know that there have been contacts with people on TikTok or videos who can bring us traffic and then we do like point agreements. But I don't think Stan has a format program for that, but we do. ⁓ Nathan Latka (17:49) Okay, where are you spending most of your time? Which product? We'll focus on that one. Stéphan Donzé (17:53) For me, I'm 100 % on Eodocs, not on Talarian. Talarian, have two separate teams. Like I said, most of the focus is on Yam and GPT for work, and Awesome Table and Form Publisher are more on the back burner, but you have two separate teams that work on the two products. Nathan Latka (17:56) AODocs. Okay, all right, so let's talk more about that. From what you've communicated so far my read is that AO docs is your your enterprise motion, right? What does the largest we don't name the customer obviously, but what's the largest customer pay you per year for AO docs? Stéphan Donzé (18:21) The top three are all above 2 million per year. Nathan Latka (18:25) wow, so you have multiple million dollar per year customers. So what are they? Stéphan Donzé (18:28) Yep. Now the top three are more, two million per year. Then I have three, four more at one plus million. And then we get down to 500, 200, and so on. But we have very big customers. Nathan Latka (18:39) This is like, this is shocking to me. So I'm looking at the website, you have customers paying you two million per year, individual customers for this, what are you upselling against? Is it number of seats, number of, like what are you upselling? Stéphan Donzé (18:44) All down. So the base, if you scroll down, you see some of the logos we can display. Google and Veolia are both in the very large category. ⁓ And the other logos are not small as well. Veolia, they are building water treatment plants. know the desalination plants that are in the Middle East at risk with the ⁓ Iran war? That's ⁓ what Veolia Water Technology does. ⁓ Nathan Latka (19:04) Well, what is Violia, so what are they using you for? Yeah. The War. Stéphan Donzé (19:21) We've recently ⁓ migrated most of their North American division from SharePoint to EODOCs. There's a very nice success story. you scroll a little up, you'll see this guy in the middle is Nicolas Stoner, he's the managing director of operations for the Middle East. And he's the one using EODOCs for ensuring that Veolia distributes the right technical documentation to the actual contractors building the water treatment plants. and ensuring that those projects are delivered on time and budget with the right profitability. So essentially at Theodox we ensure that companies use the right version of the documents, use the right authoritative information when they do their projects or they do their processes so that we reduce human errors and we accelerate the business process. We do the same with Google. We help them build data centers, which is really business critical. Nathan Latka (19:56) Interesting. Stéphan Donzé (20:19) use case for them. We are working with Airbus as well on airplane certifications. there's a lot of business critical processes running on EODOCs. And to your question, what's the metric for the pricing? Usually the number of users or the number of employees plays a role in pricing, but it's not necessarily the only metric. Sometimes we price based on the process themselves, how many documents you have to to validate or to ingest or to process every year. So it's a mix, but most of the time, number of users, number of employees is the right starting point for the pricing. Nathan Latka (21:01) It's shocking to me that Google pays you millions per year for a product, which is document management, when they have Google Sheets. I they have a whole product line around this. Why don't they build this internally? Stéphan Donzé (21:11) It's not the same because Google focuses on collaboration. When you're on Google Drive, you have your documents, I have my documents, we do whatever we want with those documents. The value proposition of Google Drive is speed ⁓ and convenience and ease of use. the ⁓ bad side, let's say, of that is nothing is controlled. You have no guarantee that what you find in my Google Drive is up to date, validated, and truthful. What AODocs does is we control those documents. We put them in a process, we have traceability, and we are able to tell you this is the authoritative version. This is the version that has to be sent to the contractor and you can build the data center from it because it's been validated by engineering, by legal, by finance, in this order. So it's a completely different value proposition. We are on control, limiting ⁓ mistakes and ensuring you don't send the wrong version of something while Google Drive is... purely on collaboration and speed. And both work together because when you use Eodox and drive together, can use, everything I say here is true for SharePoint in Microsoft, right? You have the collaboration and the control. When you're in collaboration phase, you want to go fast, but at some point you have the final version, you want to send this sales proposal to the customer, and you want to make sure you send the right price. So that's where Eodox comes into play, and we have the validation of the two. are complementaries, two different sides of the same coin. But Google, they don't have a tool for document control. It's not their business, it's not their DNA, and for this it makes sense to go ⁓ look on the market for specialists. Nathan Latka (22:46) Interesting. What's your go-to market? How are you getting new customers at AO docs? Is it outbound motion or what? Stéphan Donzé (22:52) It's outbound, it's trade shows, it's word of mouth. We get a good part of our business inbound. 50 plus percent of the sales pipeline comes from contacts on the website ⁓ or RFPs that we receive because people know we exist and they want to include us in a tender. that's ⁓ the different channels. Nathan Latka (23:18) As you test different live events like this, have you learned anything about booth construction or the sales stuff you're handing out or what you're putting on the screen to just maximize conversion at events? Stéphan Donzé (23:27) I think the one thing that works well at events is to ⁓ contact the attendees before. So typically when we do like this, this one you have here on the screen is the Microsoft Community Conference in Orlando. We were there a few weeks ago. So we do a pre-campaign on LinkedIn. Hey, we're going to be there in one month if you want to see us, blah, blah. We get the list of attendees and we contact each and every one of them and we explain our very proposition and we push to book a meeting on the booth. when we do it right, This is Google Next in Las Vegas end of April. When we do it right, we can get more traffic because people know we are there and they have added us on their list of ⁓ sponsors to visit. When it's available, we also buy speaking slots. So we have sponsor speeches, sponsor presentation, different events. there is a lot. Nathan Latka (24:20) does that cost usually? Stéphan Donzé (24:25) Small events can run in the 10 to 20,000. Big events, you can pay 100 or plus thousand dollars for. Nathan Latka (24:33) I know if it's worth it, Stefan. I mean, is it worth spending a hundred grand for a 20 minute keynote spot? Stéphan Donzé (24:37) It is, no, the keynote spot itself is an add-on on the thing, but you can pay 20 or 30K just to have a presentation. And at the big event, if the attendees are on target, our target are IT architects, ⁓ IT leaders. And if you go to a conference where those IT architects pay to come and they pay, to go there and to learn about the market and to survey what's new and to come back home with a list of new vendors that they ignored, that they didn't know exist. If you go to that conference, mean, it's 100 % on target, right? So every person you touch at this conference is potentially a future prospect. Maybe right now, maybe six months, maybe two years later. So a 100K conference for us, a single deal. from that conference ⁓ pays for it. So it looks expensive on paper and it is, but if you do it right and you come back with one or two or three leads that close from it, it can be vastly profitable. So it's a game of big deals. Nathan Latka (25:50) Last question here, many big deals? How many folks are full time Stefan at the company today? Stéphan Donzé (25:58) Today, for EODOCs, it's about 130, and the whole group, counting Talarian and the other companies, were 250. Nathan Latka (26:06) 250, wow. All right, anything that you really wanna touch on when you're going, man, this stupid podcaster, he should have asked about this thing and he didn't. Stéphan Donzé (26:14) So I think one thing we didn't discuss is the geography. We are a multicultural company. We have French and Italians and Hungarians and Americans and all kinds of It's mixing those different cultures is really interesting, very beneficial in terms of, I it sounds like cliche, but I think it's really good to have those different points of view. You have the Eastern European that see everything as very negative and are always like, is not going to work. You have the French and Italian that are more outspoken. So the mix of all this is really interesting. ⁓ beyond that, we have also a model. Sorry, I'm going to sneeze. ⁓ No. ⁓ We have a model where the engineers are in Europe. ⁓ Nathan Latka (26:50) Ha ha ha. OK, bless you ahead of time. Stéphan Donzé (27:08) the majority of the market and the customers are in the US. And ⁓ the reason I did that, first because I had my own network of engineers in France from my past at Exelit, but also because in terms of cost, in terms of ⁓ knowledge, there are very good universities, very good technical background in Europe and Eastern Europe. To build software, it's it's a good base. And then of course software, you can export it everywhere. So we take advantage of that model since 2012 and I'm very happy ⁓ about it. We have more stability and more, frankly, a better technical level with European engineers than what we could find in the US for the same budget. Nathan Latka (27:55) there you go check them out aodocs.com if you're looking for IT solution for document management or if you're looking for a job right? Alright Stefan if people want to follow you online as we wrap up here where can they find you? Stéphan Donzé (28:01) Exactly. Mostly on LinkedIn. I'm not a big Twitter guy, so if you want to follow me, ⁓ go on LinkedIn. Nathan Latka (28:14) Guys, what a story services company launched in 2007. Now today, 26, 2026 will do 60 million of revenue, 50 % of which comes from AO docs. Think of that like what Google or water treatment plants use to manage their official regulated set of documents. You know, that's 30 million of revenue under the 30 to 40 % comes from a group called Tellurian, which manages GPT for work. It's a plugin for Excel and Google sheets or yet another mail merge, a very popular Chrome extension. And lastly, a services business that'll do more than 10 million bucks this year called, I'm going get this wrong, but REV. Evol.com, rib of all, right? We love the model of services plus software. They're profitable 50 to 20 % profit margins last year where they closed 55 million bucks of revenue. 130 folks at AO docs whole group is 250. So really healthy revenue per employee as he focuses on building in the age of AI. Stefan, thank you for taking us to the top. Stéphan Donzé (29:02) Thank you for having me and bye, see you later.