From disposable AI chats
to durable macro analysis.

Stop rebuilding the same CPI, rates, and regime charts every week. Savvy turns AI analysis into reproducible, shareable artifacts—with code, data, assumptions, and outputs bound together.

If you can't rerun it, you can't trust it.

Rerunnable Audit-ready Shareable
CPI Decomposition: Base Effects vs MoM FRED Data

AI can talk macro. It can't do macro work.

Macro work compounds: you revisit the same questions across months, prints, and regimes. But AI chat results are disposable.

The session ends → the analysis disappears
Charts aren't tied to assumptions → you can't defend them
"What changed since last print?" becomes manual work again
Teammates can't reproduce your result → no trust under PM scrutiny
That's not analysis. That's a demo.

Savvy adds state to macro analysis

Not chat history. Analytical state—so every result is rerunnable.

Savvy creates an artifact that includes:

📊 Data snapshot (or query)
⚙️ Transform steps
📝 Assumptions / parameters
🐍 Executed code
📈 Charts + tables
🔍 A "how this was produced" record

Common macro workflows, done in minutes—and reusable next week

CPI print review

"Decompose CPI YoY into base effects + MoM momentum. Highlight drivers vs last print."

Chart pack Decomposition table Rerunnable artifact

Rates regime dashboard

"Plot 2s10s, term premium proxy, real yields, breakevens. Mark regime shifts since 2010."

Multi-panel chart Regime annotations Rerunnable artifact

Growth vs inflation regime map

"Create a 2x2 regime map (growth↑/↓ vs inflation↑/↓), classify last 15 years, and label key transitions."

Regime quadrant Historical classification Rerunnable artifact

Scenario table (PM-ready)

"If inflation re-accelerates by 1% and growth slows by 1%, estimate directional impacts on 2y/10y, DXY, and equities based on historical episodes."

Scenario matrix Historical analogs Rerunnable artifact

Nowcast-style tracker

"Track surprises for payrolls, CPI, PMIs vs consensus. Show rolling surprise index and correlations to rates/FX."

Surprise index Correlation matrix Rerunnable artifact

Replay prior thesis

"Re-run my 'soft landing' analysis using the latest data. What changed and what didn't?"

Updated charts Delta summary Rerunnable artifact

Speed and trust (not one or the other)

Before

  • Ask AI → get narrative
  • Copy snippets → run locally
  • Lose assumptions → can't defend
  • Rebuild charts next week

After (with Savvy)

  • Ask AI → get an artifact
  • Rerun anytime with new data
  • Share a link (not screenshots)
  • Audit assumptions and steps

How Savvy fits into your workflow

1
Connect or upload
CSV, parquet, database, API
2
Ask questions
Natural language prompts
3
Savvy executes
Generates artifacts
4
Share / rerun / extend
Across sessions and collaborators

Works with Claude Desktop via MCP protocol

Every output is packaged for reuse

📦 cpi_decomposition_jan2025.zip
  • 📄 charts.html Interactive, shareable
  • 🖼 charts.png Slides-ready
  • 🐍 analysis.py Exact code executed
  • 📊 data_snapshot.csv Or query pointer
  • ⚙️ assumptions.yml Explicit parameters
  • 📝 README.md Methodology + provenance
Same question, next week: click rerun.

Built for macro teams who get challenged

Hedge fund / asset manager macro analysts

Multi-asset strategists

Economists producing recurring research

PMs who need scenario-ready work

If your work ends with "prove it", Savvy is for you.

Make your macro analysis cumulative

Durable artifacts. Rerunnable results. Shareable trust.