For the past 40–50 years, Silicon Valley has remained the center for the formation of new technological paradigms: personal computers and internet-related startups were actively developed here, followed later by web applications and services. Today, this region has become the capital of artificial intelligence: key industry players are based here, from OpenAI and Anthropic to the research divisions of major corporations. This is where AI startups from all over the world aspire to be, to work in an environment that sets global trends.
Among those developing their project here is Ivan Cherevko. In recent years, he held the position of Chief Privacy Officer at Yandex, previously worked at Rambler, and founded Hotelscan, a hotel price comparison service. He is currently developing his own company that develops custom generative AI services.
Ivan explains his decision to develop the Maracas.ai project by the fundamental technological shift that has occurred with the rise of generative AI, which has radically altered the balance of power between corporations and small teams. While previously creating a competitive product required the resources and infrastructure of large companies, today the key work can be done by just a few specialists with the support of AI.
“There are about six people on the team with varying levels of involvement: some contribute to tasks on a one-off basis, while I collaborate with others long-term. One of my colleagues, for example, wrote a huge amount of code for our products—with her own hands, but with the help of AI agents. Our work is organized differently than in large corporations with fixed positions. It's more like the activity of an art collective or a musical band, where everyone conducts their own process. I think many companies will adopt this approach in the future," says Ivan Cherevko.
Competition in the AI market is constantly increasing. Young startups and specialists with extensive experience in team management and product creation, which Ivan undoubtedly is, are also presenting their projects.
Now, Chinese 996 culture has arrived in San Francisco: working 12 hours a day, 6 days a week, or even without days off. While more experienced developers argue about whether it's right to work around the clock without spending time with family or resting, young teams are getting ahead because they complete tasks faster. Winning this race won't be easy.
While avant-garde companies are developing world-class products, many traditional businesses continue to solve problems using old methods or implement AI inefficiently. Therefore, Ivan decided to develop his project in this direction. It helps clients automate routine operations—from marketing and customer support to internal back-office processes.
While most AI companies in the market either create a product or offer consulting services, Ivan Cherevko decided to work at the intersection of these areas and develop custom products for specific business challenges. An important element of his team's work is a scientific approach. All results are checked using quality assessment tools to not only evaluate the effectiveness of the implementation but also to continuously improve the system.
The main customers of Maracas.ai are medium and large companies with several hundred to several thousand employees, where human labor accounts for a significant volume and where optimization can yield an economically significant result. For example, for one of them, Ivan developed a customer support automation system. The created chatbots not only help process standard inquiries but also reduce customer emotional stress by showing empathy. The system resolves about 60% of inquiries independently, while the remaining 40% are passed on to operators already in a structured format.
Today, as the number of AI projects is rapidly growing, arguments about market oversaturating are being heard more and more often. But according to Ivan Cherevko, this impression is deceptive: "You can't say that business interest in artificial intelligence is declining. We are at a very early stage, and even companies with the highest level of engagement have not yet fully tapped into the potential of the technology. Yes, there's a feeling of a bubble—huge investments in data centers, in purchasing GPUs, and record market capitalizations for companies like Nvidia. All of this really reminds me of the internet history of the late 1990s, when massive investments coexisted with mass bankruptcies. But what's important is something else: the internet ultimately changed the world radically, and the same thing is happening now with artificial intelligence. We are at the dawn of a new industrial revolution.”
Ivan confirmed that some companies are disappointed with AI but emphasized that this is not due to the technology itself but rather to specific implementation errors. Many companies, riding the hype wave, invested in solutions that didn't meet real needs. As a result, expectations were not met, and management concluded that "AI doesn't work."
"Artificial intelligence isn't a magic 'make it better' button," Ivan notes. "It's a huge ecosystem of methods and tools with a very high 'skill ceiling.'" You can use them correctly or incorrectly. And unfortunately, incorrect applications are even more common than successful applications. Therefore, the failures of individual companies should not be seen as a sign of an industry crisis—it's just a stage on a long path that will take decades."
When asked what startups should be doing today, Ivan replies that the key task is not to succumb to the temptation to "relax" and offload all the work onto AI. Companies have a choice: use AI to save effort and stagnate, or to enter new markets, create more complex products, and solve problems that were previously unattainable. The temptation to relax is very strong, but it's a one-way street. The competitive environment quickly "forgets" those who stop growing. Therefore, the advice is simple: use AI tools not to relax, but to become better every day—for yourself, your employees, your clients, and the world as a whole.