Reimagining Qual: How Conveo.ai Scales Deep Insights with AI

In this episode of the CEO Series, Lenny Murphy sits down with Dieter De Mesmaeker, Co-founder & CEO of Conveo.ai, and Niels Schillewaert, Head of Research and Methodologies at Conveo.ai. This rising AI research startup is turning heads in the insights industry. Fresh off a major funding round, the team shares how they’re using AI-moderated video interviews to bring qualitative depth to scale—without sacrificing human nuance. From product innovation and market adoption to the future of multimodal AI in ethnographic research, this conversation offers a front-row look at what’s next for qual in the age of automation.

Transcript

Leonard Murphy: Hello everybody. It's Lenny Murphy with another edition of our CEO series of interviews. joined by Dieter and Neils from Conveo. Welcome gentlemen.

Dieter De Mesmaeker: Hi, nice to meet you.

Niels Schillewaert: Thank you, Lenny. Yeah.

Leonard Murphy: You as well. Neils, it is weird to say Neil's from Conveo. So, why don't we go there? Some people may be saying why. So, we'll start with you, Neils? Give your quick background and then Dieter, we'll transition into you and talking more about kind of the latest news. All right?

Niels Schillewaert: Thank you, y. And yeah, some people may know me from my past at Insights Consulting now called Human 8, where we did quite some innovations already, which I thought were transformational back then, which they were back then, but looking at what's happening today. Joined Conveo over the summer and having a blast translating everything let's say more traditional methods into the AI space. So, good to be here.

Leonard Murphy: Can't think of any better to do that, my friend. So, it's good. All right. So, Dieter?

Dieter De Mesmaeker: Yeah. Yeah. So, I'm Dieter. previously I co-founded a company called datacam.com. It's an edtech platform for data science. Grew it over about 10 years. Based out of New York, London and Belgium. And then I think about two years ago was kind of fed up doing it. The same problem for 10 years. So then back at that time AI was hugely booming. So then basically long story short I ran into Henrik, my other co-founder, and then here we are now about a year later and also having a blast.

Leonard Murphy: So for the audience, this is not our first conversation overall. I've gotten to know your team a bit over the past few months. But then we didn't really talk and then last week you released this press release that was surprising on two fronts. One, congratulations on that was certainly a significant round and a great impromater of legitimacy, right? When you pull that off. But as well as the product features that had changed quite a bit from the early version. So I don't want to steal your thunder. Why don't you talk a little bit about that because it's a great springboard for the rest of the conversation?

Dieter De Mesmaeker:  Conveo focuses on doing qual research at scale. So, as typically you have the more quant approach, very scalable but lower in depth and you have obviously the call interviews. We use AI video interviews who are moderated by our agent to bring this scale also to qual and what it makes us unique is we focus on really video interviews. So we record the full conversation from beginning to end. We make sure that we record for example the person but it could also be recording their environment recording the actions they're doing. And then we make sure that we select the highlights of those videos. If we have thematic analysis and we give a highlight reel for all those themes that we discover. So the video feature is definitely one of our unique propositions we launched, we were created last year but I think this is definitely our kind of like our marketing launch where we're going full force ahead.

Leonard Murphy: So, what's the response been? Actually, you know what? I'm gonna pick on Neil's here. So, we've been around for a long time and I'm sure like me that you thought as AI started to explode thinking, okay, this is going to be interesting and Qual is going to be the low-hanging fruit, the place where we would see the most impact. And early versions, I mean there were so many companies right out of the gate, and doing this and none of them were particularly good, right? they helped show what was possible, but they left a lot to be desired from a kind of fit for purpose standpoint. Now, here we are, I think, in a different world. So what's your take as far as how the market's reacting? What's this been like for you as helping the company find product market fit and overcoming resistance? What's your take?

Niels Schillewaert: Yeah, well, lived it in a previous life Lenny...

Leonard Murphy: Sure communities. Yeah. Yeah.

Niels Schillewaert: Communities and even just taking research online in the very early days and I felt that there was much more resistance back then I also felt that maybe back then it wasn't all working maybe we've learned a lot I don't know but what...I was skeptical to be honest when I first met these guys I had a lot of questions as a research geek and  how you do this and how you do that and how you do. But I'm amazed by the product and by the way it's designed, not to say anything bad but just about our approach. Ours is really designed based on a natural conversation and you can feel that and that's why it's really qualitative in nature. And my sense is that as I referred to earlier is that this transformation is huge. It's really massive. and it's going to put a lot of paradigms upside down. Qual and quant are going to mash. I, actually, wrote a piece about that at eight or 10 years ago. but to be honest back then I wasn't really sure how it would work and if I really believed what I wrote. But what I see it's amazing and the market response from clients is very very positive. They do accept of course that there's still things to be developed and the more you have the more features add onto it. You know how that goes. But clients are amazed by the speed, the agility, the quality of the moderation which is really an experience attribute. You have to do it before you can really believe it. We can say it, it does it, but it's really these guys the prompts that have been programmed are amazing. And then obviously that creates much depth in terms of at scale is and the speed and the reduction in cost. The other thing is that part of the market is obviously who does the interviews, the consumers, the people. And we've done satisfaction about the interviewer and it's consistently 98 plus percent that rates at a four out of a five scale or more. We have some threes here and there which then have referred to a tech type of issue or people are not really accustomed yet to the format but overall people feel very comfortable talking to the agent and even about sensitive topics like politics or socioeconomic problems or things like that. So, yeah, I'm amazed by it.

Leonard Murphy: So, I've not seen the product demo, which we should probably do that at some point soon. Last I saw it was a text prompt a few months back. So is the moderator still primarily text or have you layered in voice or are we there where we have a photorealistic avatar?

Dieter De Mesmaeker: So right now you can still read the text because especially for people who might not have the best surrounding, not the quietest environment, but you also have a voice that reads out the text for you. And we did some testing with avatars, but at least for right now, given that the interview is really free flowing, so the agent really picks into what you're saying in real time. And that's where we did notice that it's not quite there yet to really have a natural looking conversation. And people tend to be very prone to detecting if something is a little bit off in talking to an AI avatar. So, maybe at one point definitely on our radar but for now we're just focusing on making a good auditory experience 

Leonard Murphy: That makes sense. Yep.

Niels Schillewaert: And the question is Lenny I'm an innovator but I always think about what's the relevance of it so with an avatar still today what would be the added value really to have the avatar? Other than mimicking a human conversation but do we need that question mark so yeah we'll see.

Leonard Murphy: And I think there's still the whole uncanny valley, aspect of things. I mean, as amazing as some of the visual elements in AI generated content can be it's still like, that's fake. We have to stop and look at it a little bit. So, yes in a couple years that it may be radically different but I think that's very smart to not muddy the waters with the uncanny valley aspect of things.

Dieter De Mesmaeker: Yeah.

Leonard Murphy: So all right, You're coming out party. You got the seed round. You got some money to spend. You got cool people working with you like Niels. You're coming to the US to be at multiple events. What's the next year look like for you? What do you hope to accomplish? What are the goals that you've really put in place to say, "Hey, if we can do this", besides just obviously landing contracts, I know that's a piece of things, but what are the kind of the strategic goals that you have in place for the next 12 months?

Dieter De Mesmaeker: I think big focus for us right now is hiring. So, we're actually pretty small team. we're only six and obviously, six people isn't that much given all the stuff we want to do. So, we're looking to hire at least another 10, probably ending the year at about 15 to 20 people. and I think our main focus is just really increasing talent density, just finding the best people for each position. It's going to take a lot of effort and time, but I think that's the highest leverage point strategically for us to do that right now. secondly, I think I come from a background in edtech where I would talk to 10 customers and they would give me 10 different requests on what to build.  I think if you talk to our customers, it's actually quite interesting to see commonality into what they're asking for. So one of the things for example is uploading all the data. So right now convey is a platform that's interviewing getting all this information synthesized to the researcher but then they obviously also have previous studies that they've created and that's going to be one of the big focus points for us as well strategically to really become a holistic platform that incorporates new and existing research. Anything to add for you that would be a big one?

Niels Schillewaert: Multimodal AI. So now the agent is fairly adaptive. And yes, it's like a karaoke screen with a text and a voice over. That's how I compare it. And does it very very well and very adaptively. You can flip the camera. So there we go to observational research, but everything is still based on the voice. So if people don't say, "Hey, I've got Heinz ketchup in my fridge." You won't pick it up. The multimodal AI will be able to recognize objects and if people show their fridge, the agent is going to be able to say, there's a bottle of ketchup and it's a Heinz ketchup." or there's Helman's mayonnaise or there's no Tabasco or spicy sauce or anything like that. I think that's the next step in two waves even. So let's say post interview where you've got the transcript and then you have the AI analyze the video and this is for me where another big wave of transformation is going to be for ethnographical observational research is that it's going to be able to do that live and say Lenny I see you have Heinz ketchup why is that and that's going to be another uncanny valley by the way this is going to be confrontational for people that an agent a technology recognizes what people are showing, but it's definitely something that we're working on. Which is, going to open up a lot of potential use cases.

Leonard Murphy: Okay. So the plan is that we'll be at scale but individually adaptive based upon not just I'm sure already based on responses but through other stimuli that analyzing as well.

Dieter De Mesmaeker: Yeah. 

Leonard Murphy: So Are we going to call that macro IDI? What?

Niels Schillewaert: I've been trying to come up with a new word for qualiquant and observation is a third dimension on it. But yeah, maybe we should do a brainstorm or maybe it's a topic for IIEX too…

Leonard Murphy: Maybe I mean we already tackled the thing of are we market research or are we insights and analytics and Greenbook we said we're insight analytics because it's just too damn big but yes it's interesting we are bringing up whole new and qualiquant just it's clumsy So all at scale that's clumsy it. Right. Edrox communities the whole shebang. But that's the exciting part. Right. we're defining whole new categories. Now that said my opinion and you guys tell me whether you agree with this or not is that what we really are in the era of is process disruption. So the components and we still need to know what is the right business question what's the right way to get the information to answer the business question which is all that process stuff which the bulk of the industry is defined by truly and then the making sense of that and those two that beginning and end that tends to be more consultative and the middle is just more kind of field and tab and Dieter you may don't if you know that language but Niels would but even in that middle piece where technology is creating so many efficiencies there is still a service component that clients are still looking for and I always use the example of our friends of Zappi right they were the big automation platform but they still often clients wanted somebody to push the button, even if they could. So from a business model standpoint, where are you just want to be the SAS play? Are you recognizing there needs to be some level of handholding? are you looking for partners to do the front or the back end or you don't want to touch it at all? Kind of what's your thinking about that?

Dieter De Mesmaeker: I think it's a good question. I think we see two types of clients in a way. Some clients are fully self-serve literally we just give them access and they do most of the stuff themselves. Other customers I'd say that's where your team really help them be successful. And I often look at my job to kind of take his brain and put it into the AI. And it's a lot to put in there.

Leonard Murphy: I don't know. That's kind of scary. Niels and I haven't hung out a little bit. I got a sense that I don't know.

Niels Schillewaert: You don't want to put it everything in there either. Sorry, that's definitely …

Leonard Murphy: You and me both. I'm sorry, dude. Go ahead.

Dieter De Mesmaeker: But it's definitely really interesting to see Niels would talk to customers and be like hey we really need to add this and this and this or the AI need so we be better on that regard and then we take that feedback obviously really really to heart and just try and improve based on customer feedback and obviously also expertise of all those years of experience that's compounding into giving us good feedback. But we're definitely not there yet but hopefully if you can see where we're headed we haven't even been around for a year. So if you extrapo give it three four more years who knows where we end up.

Niels Schillewaert: And on that note both in the front and the back the analytics and the design in the beginning already now there is a feature of you put in your research objectives and you click on the button it'll generate a topic guide. You'll be able to chat with that you'll be able to edit You'll be able to play around with it and so that's only going to improve. So the service level is going to go down. Now the service level is there because we still have to to assist clients and don't click this button or the number of probes for example you can indicate that you don't want to make interviews too long or too repetitive you want to use visual. So these are all tips and tricks that are in a way kind of easy to AI make sure that there's hints and tips and tricks that that's automated

Dieter De Mesmaeker: Or even like your projective techniques that you often employ. But we would be like okay obviously my background is in market research so I'd be like hey do you like this product? Niels was like obviously you shouldn't be asking people directly if they like it or not. You should be using protective techniques. So when we incorporate those techniques into the AI, so then if it's creating a topic guide, it will take those learnings and create a bit more creative questions.

Leonard Murphy: Yeah, that actually brings up an interesting thought that I had not thought about til right this minute in general of what about personas of the moderators, I mean Niels, this as well. Yeah. Somebody difference between B2B or healthcare or certainly cultural so I would imagine your clients would push you to say is there especially with voice do we have a female moderator? Is there a male moderator? Is there different personas because we're talking about sensitive topics. What...That can get pretty complex pretty fast.

Niels Schillewaert: Yeah, there's some of that already. So, the languages. So, we've got American English, British English, different types of French and Arabic and all of these. So, that's one component. there's also a tone of voice that you can install. So, for a B2B audience, you want it to be more neutral business- and then you've got friendly and even humoristic, which the humoristic I say to clients don't use it because it goes a little bit over the top. So some of that you can definitely do that. There's also on the technology side different versions of languages. But I do Lenny I think the tone of voice is a good one because we've done it we've used conveyor with healthcare professionals and obviously there you don't want it to be overly friendly you want it to be neutral but other than that I don't think that that's yet at this moment something that's really crucial to our clients. I think there's still a lot of other stuff to include and the projective techniques or the quality of the moderation or not this is a big request which we've solved in the meantime or  or perfected I would say is not to make the AI agent hallucinate or say things you don't want to be said about a brand. So if you try to do that please do try it. We'll see how far you go. but the agent's been made smart enough to stop the interview or go back on track, whatever, things like that. So, those are things that are more important, I feel, than creating the personas of the moderator right now.

Leonard Murphy: Yeah. Yeah, I would agree. Although I've been playing with creating a virtual Lenny at least as just basically writing d. and trained it off of my content and it's better than I am I think overall in many ways. But that idea of training, kind of a personalized agent, if you would, and think about the broader implications of that down the road. So I hear I agree. need to overly complicate things, but I do wonder if the ultimate expression of AI is ultra personification. If we get to that point somewhere down the road, just an interesting abstract concept, which we didn't have to think about this stuff until the past couple years, did we? It was yeah. So I want to be constant your time as well as our listeners. Major lessons through all of this process right so Dieter it's your second startup Niels as well you guys have been through this whole entrepreneurial journey before things that you've learned in general and things that you have learned specifically in this endeavor that think hey I'm going to pass this on to other folks in case it's helpful.

Dieter De Mesmaeker: I think for me one of the things I really learn more now because we went through my combinator last year and basically to summarize three months of my combinator into one sentence it would build something people want. And that is I think something that at least at my previous company was a lot of friction to talk to customers especially in B2C area and so that's why I really like  Conveo initially as an idea because basically our mission is to bring that friction down to zero and make it so easy to get feedback straight from the customer and use it in a frictionless way to make better decisions. And so I really think that by focusing on really understanding the customer really well and building something that they really have a problem with that that we can solve for them and that really helps create a lot of value. And so that's I think something that we've definitely been doing right now. We've been doing a lot of partnerships with bigger companies over the last year. That's why we've kind of been into stealth mode more or less because we really wanted to perfect what we've been building focusing on really understanding their problems and how can we best solve them before we go and scale and I say right now we're ready to scale.

Niels Schillewaert: Yeah, I agree with that. And maybe one other thing which we're talking about AI and technology. and my firm believes is you won't be beaten by AI but you will be beaten by people who use AI. so the human aspect is for Conveo super important. We're not only using AI to interview real people but it's all about the people. And when I look at the team, however small and when I look at I met Dieter and Henrik before the summer and the speed at which it has accelerated and both on the business side as well as on the product side that's because you have really good talent. So my thing is also that is about people and don't compromise on that and make sure that because that creates an atmosphere of working hard but also have fun down the road and you need to be intrinsically motivated and passionate to build the best product too. So it's human driven and talent driven. So that's for me confirmed because I had that I was convinced already but I've been shown it again. 

Leonard Murphy: Cool. So, on that note, you mentioned it before, but kind of pitch to anybody looking for a gig. what roles are you looking for right now?

Dieter De Mesmaeker: We're looking for all of people in the sales side commercially, head of sales, account executives, customer success, head of marketing, head of events, cho staff, funers associate, full stack engineer. We have a few positions of those open mostly looking for the area of Belgium for more of the engineering type of positions and the other ones we're looking for New York or London is a little kind of going back and forth which the two it will be but right now we're definitely going to one of those two areas.

Leonard Murphy: And how can people reach you to be able to throw their hat in the ring?

Dieter De Mesmaeker: So if you go to Conveo.ai in the navigation bar, you should see a careers tab. If you click on it, you will go to the website of Y Combinator. There you will see all of our open jobs. You can just go post a note there, upload your resume, and we'll definitely be in touch.

Niels Schillewaert: Or if people see it as a conference or our exhibitions, they can always come up to us and talk to us. That's of course,...

Dieter De Mesmaeker: Even better.

Leonard Murphy: They get that inerson interview right out of the gate, right? So, guys, is there anything that you wanted to touch on that we have not?

Niels Schillewaert: Think this was good.

Dieter De Mesmaeker: Yeah. Yeah.

Leonard Murphy: Yeah?

Dieter De Mesmaeker: Yeah Think so.

Leonard Murphy: Congratulations again. Obviously, we already had a relationship, so I'm going glad to see that happening on a personal level. But a rising tide floats all boats and, I think that the next wave of innovators, you're part of that and that's going to drive the industry forward. So, for our listeners, that's why we do this. So, you guys have a sense of what's happening maybe behind the scenes, to push the industry forward. So, best of luck and I guess that's it. Unless you have anything else to say. Any final words?

Niels Schillewaert: All good. 

Dieter De Mesmaeker: Thanks a lot.

Niels Schillewaert: Thank you, Lenny. And see you.

Leonard Murphy: All right.

Dieter De Mesmaeker: Yes. Cheers.

Leonard Murphy: Thank you guys. Want to give a shout out to our sponsors, our editors, and most of all to our audience. Without you, we wouldn't have this excuse to talk. That's it for this edition of the CEO series. Everybody be well. Bye-bye.