How Latin American Newsrooms Turned AI Tools Into Revenue Streams

The Pattern: Solve Your Own Problem First, Then Invoice Someone Else

Argentine fact-checking outlet Chequeado built Desgrabador in 2016 to transcribe audio from YouTube videos. It was a newsroom utility, nothing more. Then they made it public. Then 200,000 people showed up every month. Then the server bills arrived.

By 2025, Chequeado had relaunched Desgrabador as a freemium product—free tier with limits, paid tier with more. The tool now covers its own costs and generates surplus revenue. Their AI suite Chequeabot followed a similar arc, now licensed to other fact-checking organizations and news outlets.

Technology monetization is still a small slice of Chequeado’s budget, but the goal is clear: get it to at least 10% of annual revenue, and use it to reduce dependence on philanthropic funding.

That’s not a vanity metric. That’s a diversification strategy.

When the Business Model Comes First

Brazilian agency Aos Fatos took a different route. Their transcription platform Escriba—launched in 2022, supporting 99 languages—was designed to make money from day one.

That’s a meaningful distinction. Chequeado stumbled into a product. Aos Fatos engineered one.

Escriba now has more than 10,000 individual and corporate subscribers. Technology revenue accounts for more than a third of Aos Fatos’ total income. Their live-transcription and claim-detection platform Busca Fatos is next in line for monetization, likely targeting other news organizations as subscribers.

The model works because the tools solve real, recurring problems—transcription, verification, speed—that exist well beyond any single newsroom.

The One-Person Shop Renting Robots

Then there’s Abraham Torres, a journalist and educator in Cancún who built an automated Instagram carousel tool for his outlet El Zenit. A student in one of his AI courses saw it and wanted something similar for her curtain shop.

Torres adapted the workflow and rented it to her for $300 a month.

He now has three clients—a curtain shop, a restaurant, and a local U.S. news outlet—all renting tools he originally built for himself. Those revenues are helping El Zenit approach sustainability less than a year after launch.

His stack: Make and n8n, both low-code or no-code platforms. No engineering team required.

Why This Is Happening Now

Two forces are converging.

First, “vibe coding”—building tools through natural-language prompts rather than traditional programming—is lowering the barrier for journalists who never touched a terminal. If you can describe what you want, you can increasingly build it.

Second, the underlying AI capabilities (transcription, summarization, content generation) have matured enough to be genuinely useful at production quality, not just demo quality.

Together, they mean a journalist with a clear problem and a few weekends can produce something worth licensing.

The Tradeoffs Worth Watching

Speed and scale come with caveats. Chequeado’s executive director put it plainly: AI amplifies output, but low-quality content is a real risk, and verification still requires significant human oversight.

That tension is especially sharp for fact-checking organizations. The same tools that accelerate journalism can accelerate misinformation if deployed carelessly. The newsrooms doing this well are the ones keeping journalists in the loop, not removing them.

There’s also a market-size question. Licensing AI tools to other small newsrooms is a real revenue stream, but it’s not a large one. The more interesting ceiling-breaker is selling to organizations outside journalism entirely—which is exactly what Torres is doing with his curtain shop client.

The Useful Takeaway

The newsrooms making this work aren’t treating AI as a cost center or a threat. They’re treating it as a product development function.

If your team is building internal tools to solve real workflow problems, the question worth asking isn’t just “does this work for us?” It’s “who else has this problem, and what would they pay to solve it?”

That’s a different kind of editorial judgment—but it might be the one that keeps the lights on.