Multiplying workforce impact: Stephanie Anani, Solutions Engineer, OpenAI

Key Concepts

• GPT 5.5: The underlying frontier model optimized with embedded financial services processes.
• OpenAI Banker Bench: A benchmarking framework used to evaluate the performance of frontier models on specific financial services tasks.
• Deep Research: An AI capability that autonomously gathers information from trusted sources to create comprehensive reports.
• Skills: Custom-configured workflows (e.g., "Three-Statement Model" or "Deck Build") that encapsulate institutional knowledge and formatting preferences.
• Extended Thinking: A feature that allows the model to process information more thoroughly to ensure higher-quality, nuanced outputs.

---

1. The Value Proposition of AI in Financial Services

Stephanie, a solution engineer at OpenAI, emphasizes that AI is designed to handle repetitive, non-judgmental tasks, allowing employees to focus on high-value decision-making.

• Impact: A recent survey indicates that 75% of ChatGPT users are now capable of performing tasks they previously could not, effectively raising the "operating level" of the workforce.
• Strategic Focus: The goal is to move beyond mere speed to increased capability, supported by five key dimensions, with a specific focus on the financial services sector.

2. Contextual Integration and Trusted Sources

To be effective in finance, AI must operate within a framework of verified data.

• App Connectors: ChatGPT integrates directly with trusted financial data providers, including Dow Jones, LSEG, and S&P.
• Excel Integration: The platform allows for natural language interaction with Excel workbooks, bridging the gap between AI intelligence and standard financial modeling tools.

3. Technical Benchmarking: OpenAI Banker Bench

OpenAI has introduced OpenAI Banker Bench to measure how frontier models handle complex financial tasks.

• Performance: According to internal testing, GPT 5.5 is currently the state-of-the-art model, outperforming other frontier models in financial-specific applications.

4. Workflow Demonstration: Investment Analysis at "Blossom Bank"

The demo illustrates a step-by-step process for an investment analyst preparing for an Investment Committee (IC) meeting:

• Step 1: Deep Research: The analyst uses the "Deep Research" capability to synthesize SEC filings, earnings transcripts, and investment presentations regarding a target company (QXO). The AI creates a structured plan, which the user can review before execution.
• Step 2: Financial Modeling: Using a custom "Three-Statement Model" skill, the AI generates an Excel workbook.
  • Transparency: The model includes comments on assumptions and sources, and all figures are backed by formulas, allowing for manual auditing.
• Step 3: Scenario Analysis: The analyst inputs specific parameters (bear, base, and bull cases). The AI processes this context to provide updated summaries and key insights in minutes rather than days.
• Step 4: Presentation Generation: Using a "Deck Build" skill combined with "Extended Thinking," the AI generates a presentation deck.
  • Output: The deck includes charts, clear decision points (e.g., "approve with conditions"), and presenter notes that explain the rationale behind the AI's conclusions, ensuring the process remains auditable.

5. Key Arguments and Perspectives

• Human-in-the-Loop: Stephanie stresses that AI does not replace human judgment; rather, it "frees up the day" so that professionals can dedicate their time to the high-level judgments that actually drive business value.
• Auditability: A recurring theme is that the AI is not a "black box." By providing formulas in Excel and rationale in presentation notes, the system ensures that financial professionals can verify and audit the AI’s work at every stage.

Synthesis

The integration of GPT 5.5 into financial workflows represents a shift from simple task automation to end-to-end process transformation. By combining trusted data connectors, institutional "skills," and rigorous benchmarking (Banker Bench), OpenAI enables financial analysts to move at the speed of the market. The ultimate takeaway is that AI acts as a force multiplier, allowing firms to transition from manual data gathering and formatting to high-level strategic analysis and decision-making.