When Fábio Seixas founded the company that is now called Softo, in 2013, his goal was to return to his roots. A technology entrepreneur since the 90s, he had spent a decade leading Camiseteria, one of the pioneers of Brazilian e-commerce, and wanted to dive back into the world of technology after this retail experience. The company was born as Jack Sistemas, became a software house along the way and gained its current name in 2019, when it underwent a rebranding and launched the DevTeam as a Service model. With the advancement of artificial intelligence, the company adjusted the route again.
Today, Softo positions itself as an applied AI solutions company. With around 60 professionals spread across Brazil, and without ever having raised external capital, the company currently serves clients such as , , and the World Health Organization (WHO). The objective is to help organizations identify opportunities to use artificial intelligence, plan solutions and put them into practice. The portfolio ranges from machine learning models for predictive analysis to autonomous agents capable of executing entire processes in the areas of marketing, finance, HR and operations.
But the most radical change in recent times is not in what Softo delivers. It’s in how she charges.
Continues after advertising
For decades, the dominant model in the software development industry was charging by the hour or by project development time, which is starting to no longer make sense with the increase in productivity brought by AI.
“The business you did in a month, you do in two days. What is the economic model that supports this commercial relationship?”, asks Fábio.
The answer that Softo found was to migrate to a model based on outcomes, or charging for results. Instead of a contract linked to the time invested, the client makes a smaller initial investment and pays additional compensation if the implemented solution actually moves the business forward: more revenue, less cost, better sales conversion.
“It’s much easier and much fairer to the client. Instead of making a large investment with an uncertain result, they invest less and the extra remuneration only comes if the result comes”, explains the founder.
Softo implemented the change just over a month ago and is already reporting greater engagement from customers and some contracts closed in the new format. For Softo, the gain is not necessarily in charging more, but in being able to serve more customers with the same team, as AI has drastically reduced development time. The goal for this year is to grow between 30% and 40%.
“The cost per functionality delivered drops by around ten times. The biggest benefit is for the customer, in the end”, says Fábio. “It’s obvious that we, in some way, take advantage of part of this value generated. This ends up being reflected in margin. But the majority of the value is passed on by the client. He is the one who will have the project completed faster, he is the one who will spend less money.”
Continues after advertising
Among the segments served by Softo are mainly . The idea is to help companies optimize processes with AI, using both providers available on the market and training open source models for the company’s specificities.
For example, Softo developed a solution for a reinsurance company that automatically analyzes contracts, compares clauses with a company standard model and alerts legal analysts about relevant discrepancies. What previously required hours of human reading can now be done automatically, with more precision and less risk of errors.
What AI has changed
To understand the new billing model, you need to understand what artificial intelligence has done to software development. In Fábio’s view, AI solved precisely the most expensive and scarce stage of the process: writing code.
Continues after advertising
“You took the most critical step of the process and solved it with AI”, he summarizes. “Before, you did 100% of a task manually. Today, AI does 90% and you adjust the remaining 10%.”
Fábio points to the Claude Code, an Anthropic tool launched in early 2025, as a game changer for the sector. For him, the turning point came with the launch of the Opus model, in November 2025. “In four months, we went from one situation to a completely different one.”
What about jobs?
The inevitable question about AI and the job market finds an optimistic answer. In Fábio’s view, if something is cheaper, more people will want to use it, not less, which generates a greater demand for professionals.
Continues after advertising
“Companies like Anthropic are increasing hiring and paying developers $400,000 a year. The job market is growing again,” he says, citing data from platforms like Indeed.
The caveat is for less experienced professionals, who may face more competition in a market that tends to prioritize the best. In the broader scenario, Fábio believes that AI will expand the universe of problems that humanity can solve, such as predictive models for climate crises, for example.
For Fábio, AI also broke down entry barriers to entrepreneurship in Brazil, allowing more people to get their ideas off the ground. “Which doesn’t mean that starting a business has become easier, because you also have more competition”, he ponders.
