Wind Alice Financial Copilot

The Problem It Solves

Financial professionals often grapple with fragmented data sources and cumbersome analytical workflows. A researcher might need to simultaneously query the Wind database, retrieve the latest research reports and announcements, integrate financial models, and manually draft reports. This process is time-consuming and prone to error. For instance, a comprehensive credit analysis requires collecting information on entity qualification, industry metrics, financial statements, and cash flow, while also comparing internal and external ratings, all before generating a structured final report.

How It Works: Core Capabilities & Key Steps

The Wind Alice Financial Copilot (alice-financial-copilot) streamlines complex financial analysis into a single call by leveraging Wind's official Alice Agent service. The host agent passes the user's natural language query verbatim as the --prompt parameter to the CLI. The backend Alice Agent automatically identifies the intent and orchestrates the appropriate financial sub-capabilities.

Its core functionalities are organized into five domains, each with specific technical modules:
1. Credit Research (Six-Dimensional Analysis): Constructs a structured credit profile covering entity qualification, industry risk, financial health, cash flow quality, rating alignment, and probability of default (PD) modeling. The system integrates Wind's risk scoring system for quantitative assessment.
2. Asset Allocation (Three-Layer Framework): Supports the entire workflow from Strategic Asset Allocation (SAA, a 3-5 year benchmark) to Tactical Asset Allocation (TAA, a 1-6 month tilts), ultimately outputting an executable list of portfolio weights under defined constraints.
3. Investment Research & Analysis: Generates "one-page investment memos" for specific stocks, conducts multi-dimensional comparisons of funds' risk-return profiles, or produces sell-side-style quarterly earnings reviews.
4. Macro & Quantitative Analysis: Interprets macro indicators like CPI and PMI, aggregates bullish/bearish views on commodity futures to compute sentiment scores, and even supports backtesting of user-defined indices.
5. Document & Data Processing: Directly generates professional research reports in Word, PPT, and Excel formats, and processes PDF documents. Its core strength, however, lies in generating content based on financial analysis, rather than generic formatting or merging.

Applicable Boundaries & Important Notes

This skill is not a general-purpose document tool. It should not be invoked for pure PDF merging/splitting, basic PowerPoint design, non-financial writing, or personal accounting. It is specialized for scenarios requiring financial data, market quotes, financial reports, or professional judgment, such as credit analysis, asset allocation, and industry research.

Key considerations when invoking:
- Analysis typically takes 2 to 15 minutes; the host agent must inform the user in advance and wait patiently.
- Complete reports may be downloaded as attachments. The host agent must only display the streamed text response from the analysis and relay the attachment path (e.g., Report saved to: ~/Downloads/xxx.pdf).
- The skill supports multi-turn conversations, but each CLI call is treated as an independent task by default. To follow up and inherit context, the host agent should automatically add the --continue-session parameter.
- On Windows, it's recommended to use the scripts/afcp.ps1 script to avoid garbled output for Chinese characters.