Finance has shifted from a monthly reporting function to a real-time operation. Sarah Friar, who joined OpenAI two years ago to build its finance team from scratch, describes the goal as "seeing the business as it changes, helping leaders act sooner, and giving finance teams more time to shape what happens next."
Friar set two ambitions for her team: a zero-day close and continuously updated automated forecasting. The zero-day close aims to give leaders a real-time, reconciled view of the company's financial position. Continuous forecasting builds on that to show how the business is changing and which decisions could alter outcomes. The team is still building toward both, but the work has already moved beyond static spreadsheets toward live tools.
Here are five lessons from that experience.
1. Give everyone access, then create a reason to use it
The first step was broad access to AI tools. Friar paired that with a finance hackathon where teams brought real work they wanted to transform. Sales engineers, for example, built IR-GPT, a custom tool grounded in approved materials for answering investor questions. The hackathon turned AI from an abstract concept into a working solution in a single day.
"You need bottom-up experimentation and top-down strategy," Friar said. "Put secure, capable AI in people's hands and let those closest to the work identify better ways of getting things done."
2. Redesign the full workflow around the decision
Finance teams spend most of their time assembling inputs: finding data, reconciling spreadsheets, building charts, and preparing presentations. AI changes the unit of work by letting teams redesign the path from source data to decision.
For the close, Friar's team is working toward a continuously reconciled view connecting approved spending plans, general-ledger actuals, purchase orders, and transaction details. AI prepares initial variance explanations and flags exceptions for human review. "The close does not disappear. What begins to disappear is the scramble to reconstruct the business after the period ends," Friar said.
3. Finance professionals become builders
Recent OpenAI research shows 40% of finance professionals' specialized AI use involves work outside traditional finance. Friar's team is proof: people who never coded are building custom AI dashboards and tools using ChatGPT Work and Codex.
One teammate supporting the advertising business built a tool that turns monthly forecasts into weekly and daily plans, accounting for weekdays and holidays. "The people who understand the problem can now shape the solution," Friar said.
Finance teams are not becoming less specialized - they are gaining the ability to carry their expertise further.
4. Pair speed with clear accountability
IR-GPT taught an important lesson about controls. The tool produces a strong first draft of investor responses in seconds, but the investor relations team reads every draft, adds judgment, and ensures consistency. "AI accelerates the work. People own the result," Friar said.
CFOs should work with IT to define which data AI systems can access, when approvals are required, and how issues are escalated. Every output should connect to a reliable source. Every forecast should carry a clear explanation.
5. Measure value per unit of intelligence
Buying more seats or using more tokens doesn't tell you much about AI's ROI. Friar recommends asking four questions for each workflow:
- Did AI complete work that mattered?
- What did it cost, including employee time, review, and rework?
- Was the result good enough to use?
- Did it help move faster or make a better decision?
The cheapest model isn't always the most economical, Friar noted. A better model can get to a reliable answer with fewer attempts and less review.
Why this matters for finance professionals
Sitting at the center of strategy, data, and risk, CFOs have a powerful mandate to lead AI transformation. The finance professionals who adopt these lessons will find themselves moving from building spreadsheets to building decision-making infrastructure. As Friar said, "You can't be what you can't see. If we want our finance teams to embrace what's possible with AI, we have to show them what it looks like."
For those interested in practical training, the AI for CFO Training learning path covers the strategies and tools described here. The AI for Finance category includes more resources on financial analysis, forecasting, and automation.
Your membership also unlocks:
