Espresso Labs featured in Valor Econômico: how AI changed developer evaluation

Aug 2026 · 6 min read

Espresso Labs was featured in Valor Econômico, Brazil's leading business newspaper, in a story about AI tool costs rivaling salaries and the new way of evaluating developers. Read the key points.

The discussion that led us to create Devint has reached Brazil's leading business newspaper. Espresso Labs is featured in a story published by Valor Econômico on August 5. For readers outside Brazil: Valor Econômico is the country's largest financial daily, roughly the Brazilian equivalent of the Financial Times or The Wall Street Journal. The headline translates as "In the AI era, tool costs already rival salaries and change how developers are evaluated".

The topic is exactly the problem we lived through in practice, and the one that led us to build a product to solve it.

The math that changed

For decades, the cost of a development team was, essentially, payroll. Tooling was a footnote: an IDE license here, a server there. AI turned that equation upside down. Between code assistant subscriptions, model token consumption and agent tooling, the monthly spend per developer stopped being a rounding error and started competing with the salary itself in technology budgets.

And when the cost per professional jumps to a new level, the next question is inevitable: how do you know that investment is coming back as productivity? Traditional evaluation metrics, lines of code, delivery counts, perceived seniority, were designed for a world where developers wrote everything by hand. They fail to capture who uses AI well, who just delegates without judgment and who produces volume without quality.

Where Espresso Labs comes into this story

We felt this distortion first inside our own operation, managing our own teams, and then confirmed it with clients: productivity per developer went up, cost per developer went up too, and performance controls fell behind. Evaluating performance became a matter of opinion, and opinion doesn't sustain business decisions.

Devint, our performance evaluation platform for technology teams, was born as the answer. It crosses metrics from the tools teams already use and summarizes each professional's performance in the DevScore, a 0 to 100 score that brings objective criteria back to management decisions: promotions, training, allocation and, now, also the return on AI spend.

We introduced Devint to the market at Rio Innovation Week 2026, one of Latin America's largest technology events, held in Rio de Janeiro, and we continue to open the first access spots for the platform.

Performance is not opinion

The Valor story confirms what we have been saying on stages and in meetings: the AI era didn't make evaluating technology teams less important, it made that evaluation harder and more expensive to get wrong. Companies that keep measuring developers with a 2020 ruler will make wrong decisions backed by numbers that look right.

Read the full story on the Valor Econômico website (in Portuguese) and, if you want to understand how Devint can bring criteria to your team's evaluation, talk to us.

Have a similar idea?

Send us a message, we will help you find the best way to bring it to reality