OpenAI: Revolutionizing Finance with AI, Five Key Lessons

Revolutionizing Finance with AI: OpenAI's Experience

A Real-Time Transformation of the Financial Function

In the world of finance, speed and efficiency have become imperatives. For OpenAI, the opportunity to integrate artificial intelligence goes far beyond accelerating account closures or updating forecasts. It is about transforming how the company perceives its evolution, enabling leaders to make decisions more quickly, and giving financial teams the time to focus on the future.

When the company integrated OpenAI two years ago, the financial team was still in its infancy, supporting a rapidly expanding business. It was necessary to build this function from the ground up, incorporating AI as a central pillar of the decision-making process and business support.

Initial Challenges and Bold Ambitions

Initially, the challenges faced were familiar: account closures and forecast updates still required a lot of manual work, involving information gathering, explaining changes, and compiling the necessary elements for decision-making. Although access to the most advanced AI tools was available, learning to redefine the financial approach around these technologies was still ongoing.

The company therefore set two ambitious goals: to achieve a zero-day close and to automate forecasts continuously. The idea behind a zero-day close is to provide leaders with a real-time, accurate, and traceable view of the company's financial situation. Continuous forecasting builds on this foundation to show the company's evolution, anticipate future events, and guide decisions that could influence outcomes.

Evolving Work Methods

Although the company is still on its way to achieving these goals, the work has already transformed operations. The limitations of static spreadsheets and manual research have been surpassed to adopt dynamic tools based on the context and comprehensive data of the company. This evolution has allowed finance professionals to create the tools necessary for their work and expand their expertise.

For the company, the true promise of an AI-native financial function lies in a team capable of understanding events in real-time, helping leaders anticipate future choices, and giving the business more time to act while outcomes can still be influenced.

Five Practical Lessons for CFOs

1. Widespread Access and Incentivizing Use

   The first step towards integrating AI was to ensure broad access for all. Employees must be able to explore AI in the context of their daily work. This access is most valuable when paired with structured experimentation around concrete problems.

   • A financial hackathon was organized, inviting business engineers to participate and transform tasks they wanted to improve. One of the outcomes was the creation of IR-GPT, a customized GPT based on documents approved by the investor relations team to answer due diligence questions.
   • This hackathon allowed the transition from AI as an abstract concept to a functional tool. In one day, participants were able to identify a recurring task, develop a solution, test it with their colleagues, and improve it.

   For CFOs, the lesson is clear: encourage bottom-up experimentation while having a clear top-down strategy. Provide secure and effective AI to employees and let those closest to the work identify better methods.

2. Redefining Workflow Around Decisions

   Financial teams spend a lot of energy gathering the necessary elements for decision-making. A review of forecasts may require searching for the latest data, reconciling spreadsheets, explaining variances, creating charts, preparing documents, and transforming them into presentations.

   AI changes this dynamic by allowing the complete journey from source data to decision to be redefined.

   • The company is working towards a different operational model. The ambition of a zero-day close is to connect approved spending plans, ledger updates, purchase orders, provisions, and transaction details into a continuously reconciled view.

3. Finance Professionals Become Creators

   The most significant transformation is that finance professionals can now build the tools they need.

   • A recent study by OpenAI revealed that 40% of the specialized use of AI by finance professionals pertains to tasks outside traditional finance, and 22% is related to engineering.

   Each team member creates dashboards and custom AI tools with ChatGPT Work and Codex. Work is evolving from static Excel models and PowerPoint presentations to live dashboards based on the context and comprehensive data of the company.

4. Balancing Speed and Control

   The experience with IR-GPT taught the importance of control. Investor due diligence questions may require an analyst to search through previous documents, draft a response, check for consistency, and coordinate the review. With a customized GPT based on approved sources, work that previously took hours can now produce a solid first draft in seconds.

   • The human role remains essential. The investor relations team reads the draft, adds judgment and context, and ensures that responses remain consistent.

5. Evaluating Value by Unit of Intelligence

   CFOs need an AI dashboard anchored in operational performance. Buying more licenses or using more tokens does not provide a complete picture. What matters is whether the work is done well and what the true cost is.

   For each workflow, ask four questions:

   • Has AI improved the quality of work?
   • Has it reduced the time needed to complete tasks?
   • Has it decreased associated costs?
   • Has it increased employee satisfaction?

These lessons demonstrate how the integration of AI in finance can transform traditional practices, offering better responsiveness and more informed decision-making.