How the Kazakh startup Amurex.ai is gaining traction in Silicon Valley

Kazakh startup Amurex.ai turned an idea into a prototype overnight and quickly attracted hundreds of users. How did they manage to do that? Read the full story below.

Arsen Kylyshbek, 19 years old, San Francisco, co-founder of The Personal Artificial Intelligence Company, Instagram

How it all started

I am originally from Astana and studied at the Republican Physics and Mathematics School. I began launching startups at the age of fourteen — with both successes and failures along the way. I have worked in various companies and countries. Two years ago, I moved to Silicon Valley, later returned to Kazakhstan, and eventually relocated to London to join the Entrepreneur First program. That is where I founded The Personal Artificial Intelligence Company, where I serve as founder and chief technology officer.

Launching the startup

Entrepreneur First brings together specialists to create startups. I joined the program in September 2024, working in a team of forty people while testing ideas and searching for a co-founder. That is how I met Sanskar Jethi, who suggested building an artificial intelligence assistant that could anticipate users’ needs.

We created a prototype on the very first night. By morning, we already had a working version, which confirmed that we were a good fit as co-founders. This experience showed that we shared similar work styles and a common vision for the project.

We began promoting the product in niche online communities. We shared information on Hacker News, Reddit, and Y Combinator forums, asking users for feedback. Overnight, our user base grew from just a few friends and family members to several hundred users, and our servers struggled to handle the load. This confirmed the strong interest in our concept and validated the direction we had chosen.

We actively engaged with users, analyzed their feedback, and continuously improved the product. Gradually, we built a community of loyal users for whom our tool became indispensable at work. One of our advisors in London once said that a successful startup should have users who are willing to defend it. We achieved that — our product inspired not just interest, but emotional attachment.

At this stage, we are not focusing on sales or partnerships. Instead, we are continuing to build a product that delivers real value and creates long-term user loyalty.

Incubation programs

Entrepreneur First is not just an accelerator — it is also a venture capital fund backed by $160 million from investors such as Reid Hoffman, founder of LinkedIn, and the Collison brothers, co-founders of Stripe.

The program lasts for twelve to thirteen weeks. The first eight weeks are dedicated to finding a co-founder, followed by about four to five weeks of product development and traction without external interference. In January, startups present their progress to the investment committee. Successful teams receive funding and continue the program in San Francisco until the end of April, focusing on the next investment round.

We received funding and are now preparing for the next round in San Francisco.

Challenges

We encountered many difficulties, but technical challenges never raised doubts for us. Regardless of the complexity, we were confident in our abilities and could quickly develop high-quality features. Some tasks required considerable time and resources, but we consistently delivered new functionality for our users.

Our main challenge emerged as our user base grew. Different user segments had different expectations, and at one point we lost focus by trying to meet all their needs. We spent several weeks developing features that were ultimately never launched because we realized they did not align with our key priorities. This led to wasted time and resources and forced us to reassess our strategy.

Fortunately, with the support of our advisors and careful analysis, we overcame the crisis. We spent up to seventeen hours a day in the office, fully immersed in work and exploring new paths forward. This experience helped us clearly define our priorities and focus on what really matters.

There is no universal advice for startup founders — every niche is unique. We are building an open-source business-to-consumer product aimed at end users, while many startups create business-to-business solutions.

The main insight we gained is that the product must be useful to the founders themselves. At some point, we realized we were losing focus and returned to the core question: why did we start this project in the first place? The answer was simple — we wanted to build something we would use ourselves.

This approach contradicts traditional startup doctrine, which suggests conducting user interviews, validating demand, and only then launching a product. That method may lead to a narrowly targeted solution, but it does not foster breakthrough innovation. Our philosophy is to experiment quickly: build a product — even in raw form — launch it, analyze feedback, and improve.

If the product is not in demand, the key is to learn from the experience and move on. Mistakes are inevitable, but it is through trial and analysis that a working model can be found. It is important to focus on creating value first, and only then on building a business.

Achievements

At this point, we have achieved significant user base growth.

We analyze our audience using three key metrics: daily active users, weekly active users, and monthly active users. Our primary focus is on weekly active users, as this metric most accurately reflects user engagement. We are seeing steady growth and an increasing number of users actively relying on our product every week.

Plans

Over the next three to five months, we will continue developing the product. At the moment, it functions as an assistant for video calls, but our goal is to make it more versatile. Video calls are a narrow vertical, so we are expanding the assistant’s capabilities to integrate into professional workflows.

Our product will support email, video conferencing, and knowledge management tools such as Notion, Obsidian, and Google Docs. We are creating an intelligent assistant that not only helps with tasks, but also remembers user preferences and provides personalized recommendations.

We have already implemented several new features and continue to run tests. Next month, we plan to begin fundraising, after which we will focus on scaling the team and accelerating growth.