📈 Claude AI works directly in Excel sidebar 📈
Users can now work with Claude in Microsoft Excel sidebar.

Anthropic launched Claude for Excel in beta, introducing a sidebar integration that enables direct AI assistance within Microsoft Excel. The tool reads, analyzes, and modifies spreadsheets with full transparency of every edit. Users can fix formulas, populate templates, or build workbooks from scratch. Anthropic introduced seven data connectors for financial services, providing Claude with real-time access to Aiera (earnings transcripts), LSEG (market data), and Moody’s (credit ratings). New finance-specific Agent Skills enable DCF modeling, company analysis, and coverage report drafting. The features are available now for Max, Enterprise, and Teams subscribers. 

🏢 Amazon cuts 14,000 corporate jobs, offering internal roles 💼
Amazon eliminates thousands of jobs driven by AI adoption and automation.

Amazon announced it will cut approximately 14,000 corporate jobs, with more cuts expected next year, in a major restructuring driven by AI adoption and automation. The layoffs affect multiple divisions including AWS, Prime Video, devices, advertising, HR, and Alexa. CEO Andy Jassy flagged the move in June, stating increased use of AI tools and agents would automate routine tasks and reduce corporate headcount. Amazon’s corporate workforce includes roughly 350,000 employees out of 1.56 million total. The company will offer affected workers 90 days to find new internal roles. This represents the deepest job cuts since 27,000 positions were eliminated in 2022-2023.  

💚 GPT-5 update tackles mental health crises 💚 
OpenAI partners with 170 clinicians to improve crisis response safety.

OpenAI released a GPT-5 update designed to better handle mental health emergencies, developed with over 170 clinicians worldwide. The model achieves 91% compliance with clinical safety standards compared to 77% for GPT-4o. It demonstrates empathy without reinforcing delusions and maintains consistent safeguards throughout extended conversations. OpenAI estimates approximately 0.07% of its 800 million weekly users show signs of psychosis or mania. This update comes amid increasing legal scrutiny to protect vulnerable users from potential harm during sensitive mental health interactions.

INTERESTING TO KNOW

📚 Musk launches Grokipedia copying Wikipedia content

Elon Musk launched Grokipedia with over 885,000 articles, but many pages copy Wikipedia content word-for-word despite promises to be a “massive improvement.” Pages for PlayStation 5 and Lincoln Mark VIII are identical to Wikipedia counterparts, displaying messages that “content is adapted from Wikipedia, licensed under Creative Commons.” Grokipedia claims Grok has fact-checked entries, though users cannot currently edit pages themselves. On controversial topics like climate change, Grokipedia suggests media and groups like Greenpeace contribute to “heightened public alarm” without “proportionate empirical evidence.”

⚠️ AI-generated fake receipts fuel expense fraud ⚠️  

Expense management platforms are reporting a dramatic surge in AI-generated fraudulent receipts, with SAP research revealing that over 70% of CFOs now believe employees are using AI to falsify expenses. AI-faked receipts currently account for up to 14% of fraudulent submissions, with some companies detecting over $1 M in falsified invoices. The fraud extends beyond receipts to creating fake businesses and realistic video calls impersonating family members. The UK has raised particular alarm about AI-powered financial scams, reporting £629 M in stolen funds from UK residents—a 17% increase in cases compared to the previous year. This represents a significant evolution in financial fraud capabilities enabled by accessible AI tools.