Tech • IA • Crypto
OpenAI on OpenAI: Stacie Faggioli, Business Finance Officer Applications, OpenAI
8/10TL;DR
OpenAI’s finance team has restructured its workflows around AI tools and agents, achieving major efficiency gains, including operating at roughly 20% of the headcount of comparable peers.
KEY POINTS
AI-native finance model
OpenAI’s finance organization has adopted an “AI-native by design” approach, fundamentally rethinking workflows rather than layering tools onto legacy processes. This includes redesigning roles, hiring strategies, and organizational structure around AI agents. The shift extends beyond internal teams to cross-functional collaboration, embedding AI into decision-making across the business.
Lean team with high output
An assessment by PwC found that the company’s finance team operates at about 20% of the size of comparable technology peers. Despite the reduced headcount, the team maintains full operational capability, demonstrating how AI-driven tooling can significantly increase productivity and reduce reliance on large teams.
Rapid deployment strategy
The finance team prioritizes early deployment and continuous iteration over waiting for fully mature tools. This approach allows employees to adapt alongside evolving AI systems, ensuring that improvements are integrated in real time and workflows remain flexible as capabilities expand.
Embedded engineering within finance
Engineers are integrated directly into the enterprise financial technology function rather than centralized in IT. This proximity to finance specialists enables faster development cycles, real-time iteration, and tools that evolve in step with operational needs, reducing bottlenecks typical of traditional software request pipelines.
Investor relations agent saves costs
A custom investor relations agent was trained on internal data and professional communication standards to respond to investor inquiries. During major fundraising rounds totaling $40 billion and $122 billion, the tool enabled rapid, consistent responses to due diligence requests. This allowed the company to manage processes internally, saving hundreds of millions of dollars in advisory fees while maintaining high-quality communication.
Scaling executive-level communication
The same investor relations agent is used beyond fundraising, including in recruiting senior executives. It helps explain equity value and company positioning with consistent, high-level messaging, effectively extending “CFO-grade” communication capabilities across teams.
AI-powered Excel modeling
Tools like ChatGPT for Excel automate complex financial modeling tasks. In one example, the system generated a full leveraged buyout model—including projections, capital structure, and recommendations—in about 10 minutes, a process that traditionally required hours or days of manual work. Outputs remain traceable and auditable, preserving financial rigor.
Codex enables non-technical automation
Codex has expanded access to software-like capabilities for non-technical staff. Finance professionals can now build dashboards, automate analyses, and generate insights without coding expertise, accelerating data-driven decision-making across functions.
Real-time marketing optimization
By feeding large datasets into Codex, the team built an ROI dashboard that analyzes marketing spend across channels, geographies, and keywords. This enables weekly reallocation of budgets toward higher-performing channels, improving efficiency and responsiveness in marketing strategy.
Sales insights from unstructured data
Codex also analyzes Gong transcripts and customer communications to track sales behavior. It identifies whether representatives are promoting new products, delivering granular insights by region, segment, and account. This allows leadership to adjust strategy mid-quarter rather than অপেক্ষing for lagging indicators.
Automated financial reporting
Complex reporting processes, such as monthly compute margin analysis, have been reduced from days of manual work to a few hours. While outputs still undergo human validation and quality assurance, automation significantly reduces repetitive effort and accelerates reporting cycles.
Operational agents reduce manual workload
AI agents now handle routine finance operations, including procurement queries, credit risk assessments, contract reviews, and vendor risk analysis. For example, a procurement agent resolves about 60% of employee inquiries, while contract review agents flag non-standard terms in bulk, improving speed and compliance.
Democratized innovation
Many of the implemented tools originated from employees closest to day-to-day problems rather than leadership directives. Broad access to AI tools has enabled grassroots innovation, with teams independently identifying and solving inefficiencies through experimentation and internal development.
CONCLUSION
OpenAI’s finance transformation illustrates how deeply integrated AI systems can redefine productivity, enabling smaller teams to operate at scale while accelerating decision-making and reducing costs.