With a team of 8 people, AI startup earns R$1 million and creates assistant for students

In the artificial intelligence market, burning cash is common, as the technological race still demands a lot of research and development. However, a Brazilian startup decided to follow the opposite path: in less than two years of operation, Nero.AI reached the R$1 million mark in revenue, growing on its own two feet and with a team of just eight people.

Now, after consolidating itself as the AI ​​“right arm” for large institutions, such as BTG Pactual, Insper and Instituto Lemann, Nero is preparing its most ambitious move: the launch of a platform based on artificial intelligence designed to democratize high performance in studies and help those who want to do well in Enem and other entrance exams.

Also read:

For the end of “AI Washing”

Nero’s trajectory began organically. Gabriel Valentim, who has a history in scientific competitions, began to be approached by friends to solve business problems using Machine Learning. The demand turned into a real business when he teamed up with economist Enrico Gazola to create an oral correction tool for students. From then on, the duo realized that the market lacked problem solvers who really understood the technology in depth, and not just companies that plugged in third-party APIs.

While many traditional software companies try to adapt their developers to the world of AI, Nero was born with research DNA. “Of our six developers, four did academic research. We already knew how to open the hood, unplug the wires and build our own engine. This gave us a tremendous advantage”, explains Gabriel Valentim.

This knowledge about algorithms allowed the company to be financially efficient. “Instead of doing very broad research to develop conversational AI for everything, we applied our research to very specific things. We have reached a point where what we need to train is so small that we don’t spend as much money, which makes us financially effective”, adds Mario Martins, mentor at Nero.AI.

Today, the company works by filtering what is a real need for what is simply called “AI Washing” — the desire of companies to use AI just to appear innovative. “If the customer arrives wanting a WhatsApp chatbot with buttons, they don’t need AI. We quickly discard what doesn’t fit and focus on where artificial intelligence really generates value”, says Valentim.

Solving real problems

Nero.AI’s portfolio quickly filled with complex use cases. For a company in the financial sector, the startup says it is automating a bureaucratic and regulated process that previously took 60 to 70 days and was expensive in working hours, delivering a simple, low-cost solution. In the healthcare sector, the team works with large databases, training models from scratch to “anticipate patients’ health status”.

One of the most prominent cases was developed for the Lemann Foundation. The institution has a network of around 800 leaders and was facing difficulties in connecting people with common interests and pains manually. Nero solved the problem by implementing a vector database, which allows searches for similar factors.

“We take all the biographies and positions and put them in this special database for AI applications. The user enters the chat and says: ‘I need a partner in the education area who has worked with the third sector and politics’. The system looks for similarities under the hood and the chat packages this in natural language”, explains Gabriel.

Edu.AI

After success in the B2B market, Nero is preparing to launch Edu.AI, its first direct product for the end consumer, at the end of April. The project, which has already won Google and Meta hackathons, has the premise of replicating the Nero team’s own study method.

Of the startup’s eight employees, six came from public schools, were medalists in knowledge olympiads and won places at the best universities in the country (such as ITA, Unicamp and USP) and even scholarships abroad.

“We interviewed these students here at home, we understood what they did differently that gave such results, and we applied their method to our technology. We created a methodology to release this to other students at a very low price”, reveals Mario Martins. The expectation is bold: to close 2026 with 15 thousand students using the platform.

Also read:

Edu.AI also innovates in its operating model. It will operate as an ultra-lean company: “the entire process of promotion, social media and audience outreach will be done using the agentification strategy (autonomous AI agents), with just three people operating everything behind it. The idea is for it to be one of the first 100% AI startups in Brazil”, projects Martins.

Diversification and accelerated growth

With no plans to seek external capital in the short term, Nero focuses on business sustainability. With the arrival of increasingly larger customers – often indicated by the current base itself – and the launch of new products, the company projects revenue growth of around 46% for the year – a target that the founders believe can be reached as early as June.

Despite the creation of off-the-shelf products, the company’s DNA should not change. “The conception of a company that is a partner with others, that researches the best way to apply AI and that combines academic innovation with marketing problems will never cease to exist, because this is the profile of the founders”, concludes Mario Martins.