AI Dependency Highlighted by Belo’s Experience: Lessons for Startups
Artificial Intelligence (AI) stands as a crucial component for contemporary startups, yet it poses risks as a potential single point of failure. What occurs when an AI tool unexpectedly ceases to function? This situation unfolded for an entrepreneur from the fintech firm Belo.
Pato Molina’s Warning on AI Reliance
Pato Molina, serving as the Chief Technology Officer (CTO) of Belo, made headlines on X recently. Over the weekend, he shared that Anthropic had abruptly suspended over 60 accounts linked to Claude, impacting Belo’s operations. This disruption affected vital workflows, employee accounts, integrations, and access to previous conversations. Molina issued a cautionary message to startups, advising against the strategy of concentrating all resources in one area.
Unexpected Account Suspension
In his post, Molina expressed his frustration regarding the abrupt suspension. He called out Claude for disabling the accounts without clear reasons and highlighted the inadequacy of the appeal process being restricted to a Google Form.
@claudeai you took down our entire organization 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.
Reasons Behind Anthropic’s Action
A screenshot shared by Molina indicated that Anthropic halted Belo’s accounts due to suspected policy violations. The notification from Anthropic stated that their automated systems identified a considerable volume of signals linked to the account that breached their Usage Policy.
The message further indicated that access to Claude was revoked. It also provided a link for users to appeal the decision.
Molina’s Concerns on AI Provider Dependency
Following the suspension, Molina expressed worry over the ramifications for companies reliant on a single AI provider, noting that the entire organisation was impacted without prior notice. He admitted to uncertainty regarding which policy might have been violated, stating they simply received an email without further explanation.
Which specific policy we breached, I haven’t the faintest idea: we simply got an email, and that was it, goodbye Claude.
The Need for Diverse AI Tools
Molina elaborated on the advantages and disadvantages of employing multiple AI platforms within a company. He noted that the primary benefit is the assurance of operational continuity during service disruptions, a concern that he is currently experiencing with Claude. While Belo also integrates Gemini, transitioning between platforms can be challenging.
Integrations, skills, conversation histories: everything lost or, in the best-case scenario, on indefinite hold. A huge lesson for any software company that relies on AI tools in critical processes. Never put all your eggs in one basket.
Trade-offs in AI Platform Utilisation
While advocating for diversification, Molina acknowledged the complexities that come with it. Training the team on multiple platforms demands both time and financial investment. Furthermore, ensuring seamless integration between different AI systems can be intricate, which compounds maintenance difficulties.
The biggest disadvantage is operational complexity: you have to train the team on each of the platforms you use, which consumes time and money. Integrations between AI platforms are also not simple, and maintenance becomes more cumbersome.