Vitalii Romanchenko, CEO & Co-founder of Elai – Interview Series

Vitalii Romanchenko, is the CEO & Co-founder of Elai, an ai video generator platform that empowers individuals to produce top-notch videos without the need for microphones, cameras, actors, or studios.

You are a serial entrepreneur having founded several start-ups, what were some of these earlier companies?

I have spent 15 years in the tech industry, working for some big companies as well as starting a couple of my own startups. In 2018, my friend Alex and I started Progressify, a mobile-first e-commerce storefront. This was an innovative approach where every e-commerce website owner could connect to our storefront and significantly increase the performance and conversion rate of their stores. At that time, this type of solution was very new, and we lacked early adopters and customers. After a year, we decided to close the company due to the fact that the market was not ready. It was the right decision, as the market indeed was not ready, and this type of solution began to rise only in 2021, around 2 years after Progressify.

What are some of your key takeaways from this experience?

During this experience, we had the opportunity to develop our skills in conducting customer development interviews and assessing market potential. We spoke with more than 200 e-commerce store owners to understand their needs and determine if our solution could be beneficial to them. Although we were unable to achieve PMF at that time, I am grateful for the lessons learned. It is essential for founders to conduct a thorough analysis before starting a company to ensure market viability and profitability.

Could you share the genesis story behind Elai?

Alex and I had a new idea after our experience with Progressify. We wanted to revolutionize video creation by making personalized videos. At that time, I was planning to create a product management training course and realized how complicated and expensive the process can be. You need a studio setup, video software, and if something goes wrong, you have to reshoot the entire course. Alex proposed that we use generative AI algorithms to create a digital clone of myself and seamlessly generate video content without the need for a camera, studio, or video editing. I was amazed by this idea, and we reached out to L&D experts to see if they needed this kind of solution. After conducting 80 customer development interviews, we received clear confirmation that there is a strong demand and market for training video generation. However, we lacked deep ML expertise to create high-quality avatars, so we teamed up with Alexey, an experienced ML lead and lecturer. This was the beginning of our R&D department, and two years later, we now have over 30+ people and one of the most advanced AI video generation solutions available on the market.

Could you discuss the different types of machine learning used to turn a prompt into video?

Certainly, first, we need to turn the text into an audio file. For this, we use a text-to-speech engine that transforms text in many languages into a high-quality audio track. Next, we need to synchronize this audio file with the lips of the digital avatar and render the video. Creating such a solution is a highly complicated process overall, and we spent 2 years in R&D research in training and creating the models to achieve top quality. For lipsync and video rendering, we use our own in-house models.

What are some of the different types of videos that can be generated?

Elai allows you to generate any type of video content, ranging from L&D videos to product explainers and personalized sales videos at scale via API. Our main use case is training videos, so the Elai video editor is built to provide the best possible experience for creating this type of content. Nevertheless, we have an advanced API as well, which many brands use to generate thousands of personalized sales or marketing videos with their custom avatars at scale to reach their customers.

What is your vision for the future of Generative AI?

Generative AI has a bright future as it tremendously increases human productivity. Regarding video generation, I believe that in 2 years, you won't be able to distinguish between real video and AI-generated content. This technology is rapidly evolving, allowing real-time conversations with digital photorealistic avatars quite soon.

You're also a Startup Mentor for Startup Wise Guys; what are some of the most interesting types of startups that you are seeing in the AI space?

I see a lot of new companies that are calling themselves AI, but in reality, they are just using the OpenAI API under the hood as their main AI component, which is fine as long as they have a good product overall. I personally prefer deep tech companies that have their own research and development team. These teams tackle complex problems, particularly those related to autonomous AI agents or advanced AI video dubbing.

What are some common traits that you see in successful startup founders?

I would say these three: curiosity, resilience, and the ability to deliver quickly. During the early stages of a startup, it's essential to identify the right product and market, learn from negative feedback, and implement improvements rapidly. Possessing these qualities will increase your chances of success, regardless of your initial idea or plan.

What do you believe will be the next big breakthrough in AI?

I truly believe in the potential of autonomous AI agents to transform our daily lives and work. Imagine having multiple intelligent AI assistants that work 24/7, empowering you to create exceptional products and possibly even guide humanity toward prosperity and success for future generations.

Thank you for the great interview, readers who wish to learn more should visit Elai.