Dialogue AI’s oversubscribed, $6 million seed round is proof that investors, researchers, and any industry that would benefit from structured feedback (generated rapidly and at scale) are excited about the possibilities of a live, AI-based conversational interviewer.
The round was led by Lightspeed Venture Partners, joined by Michael Eisner-founded Tornante Company, Seven Stars, Uncommon Projects, and angels, including the CPO of Discord, the former CPO of Nextdoor, and the former CTO of Match Group.
Dialogue AI co-founder Benjamin Lo built his product expertise at powerhouses like Apple Music, Snapchat, Farmville, and Nextdoor. Fellow co-founder Justin Hoang is a software engineer who founded and led Twitter Shops before moving on to a pre-IPO Reddit to lead monetization strategies. Rounding out the founding team, co-founder Hubert Chen leads Dialogue AI’s technical strategy as CTO.
The origins
Dialogue AI grew organically while Lo and Hoang worked on a significant redesign at Nextdoor. Growth was quickly realized in product development and shipping, but customer insights were not keeping pace.
“Not only at Nextdoor, but at our previous companies, we learned that getting research insights could take weeks,” Lo said. “We felt that AI had hit this tipping point where you could really abstract away a lot of those pain points.”
Chen, who met Hoang at Reddit, where he was an early ads engineer, and joined the team to help solve the problem.
The in-house research team brings experience from OpenAI and Airbnb.
In research, software has been the consultancy’s poor cousin
Market research is a $140 billion-per-year business, with consultancies like Gartner and McKinsey achieving $40 billion valuations. Yet software firms like Qualtrics ($12.5 billion) and Medallia ($6.4 billion) are worth a fraction of the consultancies.
Why is the sector behind? In short, research doesn’t scale well.
“In a pre-AI world, you had this idea of qualitative research, this idea of sitting down with your customer, interviewing them, getting really in-depth insights,” Lo explained. “Because of that, you could only interview a (few) people, and you were typically at companies that complement this with quantitative research, surveys at scale. In an AI world, you see this trend where you can blend qualitative and quantitative research together to get those really rich conversational insights at basically the speed and scale of a survey.
“We can democratize market research, where it’s accessible to everyone. Whether you are a salesperson, marketer, designer, engineer, or journalist, we help people create, practice research studies where they input their high-level learning objectives, and then we conduct practice research studies for you. That’s also part of our vision.”
Dialogue AI’s tech works by sourcing participants and creating best practices for research studies. Provide the AI chatbot with learning objectives, and it produces a best-practice research plan. Then, the AI interviewer runs hundreds of interviews, with insights synthesized in real-time.
Wayfair is an early Dialogue AI customer, using the tech to understand customer flows and, critically, how can they improve purchase discovery so shoppers can easily compare alternatives and be more likely to complete checkout?
Dialogue AI conducted usability testing at scale to quickly generate insights and make product improvements to better personalize the shopping experience. And those insights were generated within hours. Lo said that alleviates the pressure felt by small research teams to satisfy multiple stakeholders.
The Next AI Frontier
Is society ready for lengthy conversations with AI researchers? Lo, Hoang, and Chen spent significant time prioritizing a model that closely mimics an expert human researcher. Hoang said Dialogue AI’s timing is perfect, as Talk AI is an emerging trend and foundational AI agents are feeling more natural every quarter.
“People are more honest and blunt with an AI interviewer versus a human moderator,” Lo added. “And when you think about it, it makes sense, right? When you’re talking to a person, you may be honest, but you’ll try and maybe sugarcoat things a little bit. People are really unfiltered when they talk to an AI interviewer; they really give it straight. From an insight standpoint, that’s actually what you want, right?”
“This is a really interesting trend, where you actually extract better insights, we think, using an AI interviewing method versus previously. No matter how skilled or neutral the human moderator is just but the fact, by being a person, there is some inherent bias.”
Investors and research communities are excited about the prospects of improvements to longitudinal studies. Input business contexts and product requirements up front, and consistent output is generated throughout the study. As the customer conducts more studies, context is enriched, and insights become more powerful. E-commerce sites can analyze feedback and interview data from repeat buyers and those who don’t buy, to learn what works and what does not.
Looking ahead
When Lo, Hoang, and Chen met with VCs, many shared their excitement over their vision for Dialogue AI, the founders’ experience, and the well-known brands already on board.
Looking ahead, Dialogue AI is working to ensure that its technology can meet customers wherever they are. That means expanding across platforms and testing new types of creatives.
“We are also building out what we’re calling our research repository, where we’ll build a platform that allows our platform to continuously build that context over time, and then allow people to query that data whenever they want, across all of their studies,” Lo concluded. “Those are two big areas that we’re focused on.”
