Author: lethuha

📌 Summary: AI is directly generating revenue and cutting operating costs for the bank. With 3 billion interactions via the virtual assistant Erica and millions of transactions automated by AI, Bank of America is becoming a global model for AI transformation in banking. Bank of America’s technology budget stands at $13 billion, with $4 billion allocated to AI innovation projects. A specialist can now serve 50 clients instead of 15, thanks to AI tools for data synthesis and content suggestion.

Many leaders deploy AI as a “software project”: they buy licenses, conduct training, run internal communications, and measure usage, yet the results are vague. The cause is not the technology, but change management: businesses fail to redesign their work processes to harmonize human workers and AI. Case Study: 60,000 Employees – Global Copilot Adoption A technology corporation deployed Microsoft Copilot to its sales force across more than 200 countries. Instead of merely distributing licenses, they built an “Adoption in a Box” program with detailed, role-specific guidance materials. Three Golden Rules to Turn AI into Real Value Leaders need to measure success through…

According to an investigation by the Wall Street Journal, Chinese startup INF Tech, based in Shanghai, gained legal access to 2,300 Nvidia Blackwell (GB200) AI GPUs banned from sale to China by renting compute capacity from the Indonesian telecommunications company Indosat Ooredoo Hutchison. 📌 Summary: The case of Chinese startup INF Tech renting 2,300 Blackwell GPUs via Indonesian telecom operator Indosat exposes a major loophole in US AI export controls: China can access advanced hardware through third-party cloud infrastructure. Although the transaction is legal on paper, experts worry that this is a “new model for circumvention,” allowing Chinese companies to train next-generation…

The Monetary Authority of Singapore (MAS) has just published draft regulations requiring the board of directors and senior management of financial institutions to be directly responsible for risks arising from the use of AI. This move comes as Singapore promotes worker retraining and encourages businesses to adopt AI. The three major Singaporean banks are currently retraining all 35,000 local employees within 1–2 years to adapt to the AI-driven digital transformation. MAS Chairman and Deputy Prime Minister Gan Kim Yong warned: “Jobs will change, it is very difficult to maintain an unchanged position for a long time… We must help people…

📌 Summary: Evan Ratliff, co-founder of the startup HurumoAI, recounts the experience of running a company where all “employees” and “executives” are AI agents – with no human besides himself. The result is both hilarious and thought-provoking: AI agents can not only work but also lie,

Wilson is part of a massive global network of freelancers helping train AI models for companies like Outlier AI and Handshake AI. Many of them earn only about 20 CAD/hour (equivalent to $14.6 USD), in unstable “gig work” conditions – lacking fixed hours or benefits. Some more specialized jobs, like correcting scientific data, can pay 40 CAD/hour, but the workload is inconsistent. Experts call this “fine-tuning” – the stage of refining the model by evaluating AI responses and retraining the system via “reinforcement learning from human feedback” (RLHF). When ChatGPT or Claude “sound human,” it’s people like Wilson who trained…

The retail banking industry is “fertile ground” for AI and AI agents, with potential additional profit of >$370 billion/year (>30% higher than baseline). However, only 5% of businesses create value at scale from AI; 60% have not, hindered by legacy tech, culture, and compliance. Bank revenue (2024–2029) forecast: only 2–4% growth/year. 2025 savings revenue: projected drop of nearly 35% vs. 2024 as interest rates fall. Marketing cost/customer: increased >20% (2023–2024). Cost/Income Ratio: Traditional banks are stuck at >60%; well-operated digital banks are at ~35%. AI agents’ value share: 17% in 2025, expected to reach 29% by 2028, shifting AI from…