Belgian AI startup Conveo lands $50M and challenges how brands read your mind

Conveo, a two-year-old artificial intelligence startup founded in Belgium, has closed a $50 million Series A funding round led by DST Global Partners, Balderton Capital, Visionaries, 6 Degrees Capital, and Y Combinator. The September round brings total funding to $55.8 million and has propelled the company’s workforce from 15 to 80 employees. As part of its expansion, Conveo has opened offices in New York, San Francisco, and London.

Co-founder Hendrik Van Hove relocated from Brussels to Manhattan to lead the North American push. He cited the density of executive leadership in the U.S. commercial market. “Today I have a meeting with the CEO of a protein brand, an hour later, I go to a marketing event where there will be CMOs of Fortune 500 companies, and then I have a dinner with one of the largest pharma agency owners in the U.S. In Belgium, you’d have to take a flight and plan two weeks in advance to have even one of those meetings,” Van Hove said.

Despite moving commercial operations overseas, Van Hove emphasized that Conveo maintains its primary engineering and product team in Europe. He noted low workforce churn and strong employee retention compared to major U.S. tech hubs.

The concept for Conveo’s platform originated during Van Hove’s tenure as a management consultant at McKinsey & Company, the global consulting firm founded in 1926. While analyzing survey data and conducting late-night research interviews for a multinational pharmaceutical client, Van Hove identified consumer interview analysis as a prime application for emerging large language models. After partnering with co-founder Dieter De Mesmaeker, the startup was accepted into Y Combinator, the San Francisco-based accelerator founded in 2005 that previously launched companies including Airbnb, Stripe, and DoorDash.

Conveo’s core software, launched in 2024, uses conversational AI agents to conduct and analyze automated video interviews with consumers. The company’s technology is designed to bridge a historical divide in market research between quantitative methods—which rely on large-scale numerical surveys and statistical sampling—and qualitative methods, such as in-depth focus groups and ethnographic studies, which seek to understand consumer motivations through open-ended conversation.

The platform allows corporate product and marketing teams to run hundreds or thousands of qualitative video interviews simultaneously. The AI moderator analyzes video feeds in real time, assessing participant responses, body language, and vocal tone to detect recurring patterns, unmet consumer needs, and emerging market trends.

“If you are doing a thousand interviews a month, you can start asking questions like: What is changing? What are people talking about? What are the emerging trends? What are the things that we don’t know that we don’t know?” Van Hove said. “And then you can start using that data to inform your decision making on a continuous basis.”

Van Hove highlighted the platform’s operational efficiency in cross-border research, such as European corporations conducting consumer studies in Asian markets. “You don’t have one interviewer per market; it’s the same AI analyzing every market,” Van Hove said. “There are no translation issues. The data is perfectly captured, so you can go back to it months later. That’s when the ROI is clear.”

The rapid adoption of AI moderation has generated friction within the global market research sector, an industry estimated to generate tens of billions of dollars annually. Traditional qualitative research methodologies were developed in the mid-20th century, notably through the pioneering focus group work of sociologists like Robert K. Merton during World War II, relying heavily on human intuition, nuance, and interpersonal trust.

Simon Shaw, director of behavioral science at the market research consultancy Trinity McQueen and a board member of the Association of Qualitative Research (AQR), cautioned that AI tools risk flattening the analytical process. “I see it as a supplement, not a substitute for human-to-human research,” Shaw said. “These platforms are fast and cost-efficient… but I think there are real differences. Good qualitative research is relational, reflective, and nonlinear. There are moments of inspiration, which is a very human thing.”

“There’s no time constraint; there’s no human judgement,” Van Hove argued, adding that automated moderators eliminate human interviewer bias and social desirability bias, where human subjects alter their opinions to avoid pleasing or offending a live researcher. “People are typically not willing to share negative feedback when there’s a human interviewer, but they’ll be honest with AI.”

Shaw added that AI moderation acts primarily as a tool for “extracting information, rather than having a relational exchange with somebody,” limiting its capacity to build genuine rapport or uncover subtle cultural nuances.

Van Hove acknowledged the structural limits of automated software, stating that Conveo is not intended to replace physical human observation entirely. “We’re not trying to be better than a human, it’s just not possible,” Van Hove said. “When you’re in someone’s house, seeing how they live, there are all these other factors you can add [to an analysis of their behavior, which AI cannot].”

The debate over automated research tools touches on long-standing questions regarding how companies gather consumer intelligence to guide product development. In May 1998, as Apple Inc. prepared to launch the translucent Bondi Blue iMac G3—a product that marked the company’s return to profitability following co-founder Steve Jobs’ return as CEO in 1997—Jobs addressed the limits of structured customer feedback in an interview with Businessweek. While noting that Apple tracked industry data and existing customer metrics, Jobs expressed skepticism toward relying on consumer focus groups to invent breakthrough hardware. “A lot of times, people don’t know what they want until you show it to them,” Jobs noted at the time.

Supporters of AI qualitative platforms argue that automated analysis of large interview datasets allows companies to identify those unarticulated consumer preferences earlier in the development cycle without relying solely on executive intuition or static numerical surveys.

As Conveo expands its operations across Europe and the U.S., its founders view the current footprint as an initial phase of broader deployment across corporate decision-making workflows. “We are only at 1% of our journey, and so I feel like there’s still so much ahead of us,” Van Hove said. “There’s so much that we still need to get out of the ecosystem, in Europe and in the U.S.”