Anthropic just brought Claude into Excel

Hello Readers👀 

Curious about the biggest AI moves today? You’re in the right place. In today’s edition, we have:

💼 Claude moves into finance with Excel and live data integrations

⚡️From Idea to Video in Seconds: Meet Seedance Fast on Freepik

🧩 GPT-5 adds deeper empathy and safety for mental health users

🔨Tools and Shifts you cannot miss

💼 Claude moves into finance with Excel and live data integrations

Anthropic is taking a major step into financial services with Claude for Excel, new live data connectors, and purpose-built Agent Skills. The update allows analysts to work directly with Claude in spreadsheets, access real-time market data, and automate complex financial models.

The Shift:

1. Claude for Excel launches in beta: The Excel sidebar lets users chat with Claude to explain formulas, fix errors, populate templates, or build entire models. Every edit is transparent; users can see exactly which cells were changed and why.

2. Real-time finance data: New connectors link Claude to platforms like Aiera, LSEG, and Moody’s for live earnings calls, credit ratings, and market metrics. This integration makes Claude more reliable for daily financial analysis and reporting.

3. Finance-specific Agent Skills: Claude now includes pre-built Skills for cash flow modeling, company analysis, and due diligence summaries. These tasks run directly within Excel or Claude’s apps, simplifying workflows across teams.

4. Benchmark-topping performance: Backed by Sonnet 4.5, Claude achieved 55.3% accuracy on the Finance Agent benchmark, outperforming peers on structured financial tasks and analysis quality.

Claude’s expansion into finance blends AI reasoning with real-world data access, finally bridging the gap between modeling and decision-making. For analysts and finance teams, it’s a move toward faster insights, fewer manual errors, and a future where AI becomes a trusted co-analyst inside Excel.

^{TOGETHER WITH PROTON}

Free email without sacrificing your privacy

Gmail is free, but you pay with your data. Proton Mail is different.

We don’t scan your messages. We don’t sell your behavior. We don’t follow you across the internet.

Proton Mail gives you full-featured, private email without surveillance or creepy profiling. It’s email that respects your time, your attention, and your boundaries.

Email doesn’t have to cost your privacy.

Ditch the Gmail data grab

⚡️From Idea to Video in Seconds: Meet Seedance Fast on Freepik

Video production just got a massive speed boost.

Freepik now lets you generate high-quality, editable video variations in seconds using Seedance Fast, ByteDance’s fastest AI video model yet.

How It Works

1. Go to Freepik AI Video

2. Create an image or upload one

3. Click “Animate” with Seedance Fast for instant results

You can now scale product visuals, test multiple ad variants, or craft launch videos instantly, without giving up creative control.

Perfect for rapid testing, storytelling, and dynamic DTC campaigns.

Try it free on Freepik and generate your next scroll-stopper in seconds.

🧩 GPT-5 adds deeper empathy and safety for mental health users

OpenAI has introduced major GPT-5 safety upgrades focused on handling mental health emergencies more responsibly. Built with input from 170 mental health experts, the system aims to better recognize distress, guide users toward real-world help, and respond with empathy without reinforcing harmful beliefs.

The Shift:

1. Expert-led redesign: The update was shaped by a global network of psychiatrists and psychologists who trained GPT-5 to handle crises like psychosis, mania, or suicidal ideation more safely. The model now redirects users to professional support when needed.

GPT-5 adds new detection systems for distress and emotional over-reliance while improving stability in long chats. OpenAI also expanded hotline access and re-routed sensitive conversations to safer models.

2. Measured progress: Clinical evaluations show GPT-5 performs 91% compliant with mental health guidelines, up from 77% on GPT-4o. Experts found 39–52% fewer unsafe responses across all major mental health scenarios.

Roughly 0.07% of users display high-risk signals, making this a vital yet complex challenge for OpenAI. The company faces ongoing legal scrutiny but continues to emphasize empathy grounded in reality, not illusion.

As AI becomes a daily confidant for millions, its ability to respond safely in moments of crisis could save lives, or cause harm if mishandled. GPT-5’s updates mark real progress toward ethical AI companionship, but they also highlight how fragile and high-stakes this responsibility remains.

🤳AI Nugget of the Day

🔨AI Tools for the Shift

🧩 VersunoAI – The all-in-one platform for unified AI asset management that helps individuals and teams stay organized, focused, and productive.

🖼 Image Upscale by Picsart – Automatically enhance and resize user-uploaded images without any loss in photo quality.

⏰ mattrAlarm – An AI assistant that literally calls to wake you up and remind you to get things done.

📣 SocialRails – Create, schedule, and publish content across 9 major platforms, saving over 20 hours a month.

🌍 Google Earth AI – Built with Gemini’s reasoning, it helps enterprises and governments monitor the environment and respond to disasters faster.

🚀Quick Shifts

🎬 Adobe introduced new AI audio tools for creators, Generate Soundtrack and Generate Speech, inside its redesigned Firefly app. The tools can automatically create synced background music and multilingual voiceovers using licensed, copyright-safe data.

⚡ AMD and the U.S. Department of Energy announced a $1B partnership to build two advanced supercomputers, aimed at accelerating innovation in energy, medicine, and national security research.

💸 Expense management platforms report a surge in AI-generated receipt fraud, with up to 14% of submissions now falsified, totaling over $1M in detected fake invoices across companies.

🚀 Qualcomm launched its new AI200 data center chip, targeting Nvidia’s dominance in AI acceleration, with Saudi startup Humain becoming the first customer for the next-generation AI200 accelerators.

🛍️ PayPal partnered with OpenAI to enable in-app shopping payments within ChatGPT by 2026, using the Agentic Commerce Protocol and Instant Checkout, letting users complete purchases directly via PayPal wallets.

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇