When Mika writes about paradigm shifts - about recognizing when optimization stops being enough - he's describing something happening right now in one of the most consequential business tools ever built: Excel.
Anthropic just embedded Claude directly inside Microsoft Excel. Not as a separate app you paste data into. As a native sidebar with complete awareness of your workbook - every tab, every formula, every cell reference.
In the video below, Nate B. Jones built an 11-tab rent-vs-buy financial model - with sensitivity analysis, opportunity cost comparisons against S&P 500 returns, and dynamic tax calculations across every U.S. zip code - in 10 minutes. The time it takes to make a good cup of coffee.
This isn't incremental improvement. This is what Mika calls above a strategic inflection point.
Claude doesn't just help write formulas (any AI can do that now). It reasons about the entire structure of multi-tab workbooks. When the context window maxed out mid-build, Claude examined the existing tabs, inferred what was missing, and continued building without explicit instructions. That's the Opus 4.5 model understanding architecture, not just executing commands.
Norway's sovereign wealth fund has been using Claude in Excel during the enterprise beta. Their estimate: 213,000 hours saved. That's real institutional proof, not a demo.
The data partnerships. Through MCP (Model Context Protocol - see issue 12), Claude pulls from LSEG for live market data, Moody's for credit ratings across 600 million entities, S&P Capital IQ for financials, FactSet and Morningstar for research, Pitchbook for private company intelligence. Any language model can help you write a SUMIF formula - that's table stakes. What a generic model can't do is pull this morning's pricing from the London Stock Exchange, cross-reference against current Moody's credit ratings, check fundamentals from S&P, and update your comparable company analysis in one workflow.
The change trail. Every AI action gets logged transparently. This matters because finance models get audited, reviewed by skeptical colleagues, handed to successors who need to understand logic. The ability to demonstrate how AI-assisted changes were made is the difference between a tool you can deploy and a liability compliance will never approve.
Local file support. Microsoft Copilot for Excel requires files saved to OneDrive with autosave enabled. Many finance teams hate this - they want control over when work gets saved. Claude works with local files. A product decision that meets users where they are.
Claude struggles with certain specialized data that isn't easily publicly available - benchmark datasets and niche research may require manual pasting. Charting is functional but not beautiful; if you want truly polished visualizations, expect five minutes of formatting work. Context windows on complex builds can max out, requiring you to clear the chat and ask Claude to infer the build plan from existing tabs. These are errors you can live with when you've compressed weeks of work into minutes, but worth knowing.
Claude for Excel is available now on the Pro tier ($20/month). The add-in installs as a native sidebar inside Excel. For finance leaders asking Mika's foundational questions - where will AI change how we create value? - this is a concrete place to start experimenting.