Multiplying workforce impact: Stephanie Anani, Solutions Engineer, OpenAI

### TL;DR OpenAI Solutions Engineer Stephanie Anani demonstrates how ChatGPT frees human judgment in financial services from repetitive tasks — through context, deep research, Excel financial modeling, and presentation generation — compressing investment analysis from hours to minutes. --- ## From Individual to Team AI Empowerment Employees spend hours each day on tasks that "need to be done but don't require human judgment." AI's real value lies in freeing that time so people can focus on core decisions that create enterprise value. A recent OpenAI survey found that 75% of ChatGPT users can accomplish things they couldn't before — not just efficiency gains, but an overall leap in how they operate. To help employees work with this augmented capability, OpenAI provides support across five key dimensions. Today's focus: why ChatGPT is the preferred platform for financial services. --- ## Key Pillars: Context, Connectivity & Model Capability ### Trusted Information Sources - **App connectors**: Already integrated with the most trusted sources in financial services — Dow Jones, LSEG, S&P — and expanding continuously. - **Native Excel integration**: Allows users to interact with workbooks using natural language, directly within the industry's most common tool. ### Model Upgrade: GPT 5.5 Underpinning these capabilities is GPT 5.5. To validate model performance in financial services scenarios, OpenAI developed the **OpenAI Banker Bench** — a benchmark quantifying how frontier models perform on financial tasks. Results show GPT 5.5 outperforms all other frontier models. --- ## Demo: A Day in the Life of an Investment Analyst Assume I'm an investment analyst at Blossom Bank. My manager wants a fresh view on QXO for tomorrow's Investment Committee (IC) meeting. Here's the complete workflow in ChatGPT. ### Step 1: Deep Research 1. **Enable Deep Research**: Activate ChatGPT's "deep research" feature to get comprehensive information on QXO. 2. **Connect trusted data sources**: Ensure analysis is based on authoritative sources like SEC filings, earnings call transcripts, and investor presentation materials. 3. **Set the prompt**: Request creation of a deep investment profile, emphasizing SEC filings, earnings call transcripts, and investor presentation materials. ChatGPT automatically formulates a research plan; the user can review and adjust. After confirmation, the report is generated. **Key finding**: In the generated report, I notice QXO is acquiring Kodiak. This will significantly impact investment decisions. ### Step 2: Excel Financial Model How to turn information into concrete numbers? Use ChatGPT to create an Excel workbook for the investment statements. - **Skill reuse**: A pre-defined skill — "Blossom Bank's three-statement model" — contains specific construction methods. No need for detailed prompts; just a simple description of what's needed. - **Transparent & auditable**: The generated workbook is not a black box — all assumptions and sources are annotated, every number has a formula, and users can step through each calculation. ### Step 3: Scenario Analysis After discussing with the manager, we decide to quantify three scenarios for QXO: **Pessimistic, Base, Optimistic**. Feed all the background context into ChatGPT and Excel; the system reads the context and executes. The output includes a summary of changes, key numbers, and actionable insights directly usable for analysis. **Impact**: What used to take hours or days now takes minutes. The team can respond almost in sync with the market. ### Step 4: Investment Committee Presentation Turn the analysis into a presentation that directly supports decision-making. 1. **Use slide generation skill**: Preloaded with Blossom Bank's preferred slide creation methodology. 2. **Enable extended thinking**: Let ChatGPT spend more time ensuring the output matches expectations. 3. **Include all context**: The deep research report and valuation analysis. Generated output: - Clear decision point: **Conditional approval**. - Includes **charts** that clearly display decision points and valuation results. - **Presenter notes** contain decision rationale annotations, making it easy to review the entire process from start to finish. --- ## Core Takeaway: AI Doesn't Replace Judgment — It Amplifies It From trusted information sources → deep research → financial model → decision point, the entire workflow is seamlessly connected. AI hasn't replaced the need for human judgment; it's simply freed up time so you can spend energy where judgment truly matters. --- **Source**: Multiplying workforce impact: Stephanie Anani, Solutions Engineer, OpenAI - YouTube (https://www.youtube.com/watch?v=m2TV8slGQKc)