Anthropic abruptly suspended more than 60 Claude AI accounts belonging to Belo, a Latin American fintech app focused on international transfers, currency exchange, and prepaid cards, on 18 April. Belo CTO Pato Molina pressed the issue in a post on X, saying the sudden shutdown temporarily crippled the startup’s internal workflows. Molina also highlighted the risks of companies relying on a single AI provider for critical functions.
In his first post on X, Molina wrote, “@claudeai you took down our entire organisation with 60+ accounts belonging to a legitimate company for no apparent reason, without any explanations. The only way to appeal the decision is by filling out a Google Form? Very bad UX and customer service.”
He also attached a screenshot of the suspension notice from Anthropic’s Safeguards Team, which cited “a high volume of signals associated with your account which violate our Usage Policy” without specifying what those signals or violations were.
Why did Anthropic take down 60 Claude accounts?
Molina, in a follow-up post in Spanish, said he had “no idea what specific policy we violated: we simply received an email and that was it.”
“Suddenly, more than 60 people were left without a fundamental tool for getting work done. Integrations, skills, conversation histories: everything lost or, in the best-case scenario, on indefinite hold,” he added.
“If you want to appeal the decision, you have to fill out a Google Form, as ridiculous as that sounds,” he wrote.
Molina also took the suspension as a major lesson for any software company relying on AI tools for critical functions: ‘Never put all your eggs in one basket.’
In a follow-up post, the CTO explained the advantages and disadvantages of using multiple AI platforms within a company; the main advantage is operational continuity during outages, but switching between models is not easy either, he wrote. Molina said that Belo also has access to Gemini, but switching to Google’s AI would mean losing conversation history and integrated processes.
He added that many companies end up “marrying” certain providers with a good service track record. Molina added that he would eventually switch to OpenAI’s Codex, soon.
“If we don’t solve the Claude issue soon, we’ll switch to Codex for development,” wrote the Belo CTO.
Anthropic reinstates account
More than 15 hours after his initial post went viral, Molina announced that Anthropic had finally restored access to the company’s Claude accounts.
“They restored our access. Apparently it was a false positive. And as always: Twitter is a service,” he wrote.
He also attached an email from Anthropic’s Safeguards Team, which stated that the company had investigated the appeal and apologised for the inconvenience.
