Interview, Fireside Chat
How AI Is Turning Figma’s Finance Team Into Builders
- Role Evolution & Strategic Shift: The CFO role is shifting from a "traffic cop" or operational leader focused on data reconciliation to a "business enabler" and "change agent" who designs automation systems and empowers teams to act as business owners.
- Hiring Criteria: Praveer Malwani prioritizes "extremely intellectually curious" candidates who can solve for business outcomes, possess data fluency, and structure their thoughts to make sense of data while avoiding the outsourcing of critical thinking.
- AI as a Tool, Not a Replacement: AI is viewed as an empowerment tool to handle tedious tasks, but leaders must remain the "thinker" to ensure the model output is understood, sensitive to change, and not "slop" (low-quality, high-volume output).
- Efficiency Gains: Automation has reduced data reconciliation and anomaly detection timelines from weeks (or days of SQL waiting) to approximately one hour, allowing for daily or even hourly metric reviews rather than monthly or quarterly cycles.
- Specific Automation Use Cases at Figma:
- Credit & Collections: Rebuilt an entire workflow for invoice tracking, cash collection, and NetSuite reconciliation, freeing the team to focus on high-impact receivable reduction strategies.
- Corporate Development: Utilized "vibe coding" (generative AI) to prototype products for diligence, enabling finance professionals to assess technical moats and product viability without relying solely on the product team.
- Investment Memos: Created "leader personas" to simulate investment committee feedback (e.g., "What would Prevere say?"), reducing cycle times and identifying common feedback themes before human review.
- Travel & Expense (T&E): Stitched fuzzy data from Navan, SIP, and Ramp to suggest policy changes, saving the team from manual cross-referencing and identifying directionally correct adjustments.
- Organizational Structure & "Finance Engineers": A "messy middle" exists where decentralized builders (finance engineers) create initial prototypes, which ideally need centralization by a modern CIO/Systems team to ensure scalability, security, and cross-org sharing.
- Data Integrity as the Primary Challenge: 95% of the challenge in AI integration is "getting your data in the right spot"; without consistent definitions and clean upstream data, AI models cannot produce trustworthy outputs.
- Future Headcount Outlook: If business complexity remains static, the finance team may not need the same number of people a year from now due to automation; however, Figma's complexity is currently outpacing efficiency gains.
- Tech Stack Strategy: A hybrid "build vs. buy" approach is used; off-the-shelf tools are preferred for core systems of record, while "build" is prioritized for workflow automation and specific use cases using tools like Codex, ChatGPT, and Figma Make.
- Recruiting Advantage: Access to advanced AI tools and the ability to work in an environment of continuous learning and "maker" culture are becoming significant differentiators in attracting top finance talent.
- Forward-Looking Statements (12-18 Months):
- Automation Depth: Expectation that more deterministic processes will become fully automated, though audit-ready outputs for business-critical systems remain a high bar.
- Talent Development: Focus on hiring leaders who operate with the same level of "business ownership" as CEOs, prioritizing influence, communication, and the ability to simplify business models for the organization.
- Skill Adaptation: Core fundamentals (analytical thinking, Excel, accounting principles) remain essential, but the "automation first" mindset and ability to iterate on tools are becoming the new baseline requirements.
- Advice for Startups: Begin with "data integrity" and rigorous metric definitions before attempting automation; without a solid foundation of tracking and truth, automation will only scale errors.
- Cultural Shift: A "step function change" has occurred over the last nine months, with finance teams moving from using AI for drafting documents to driving core data reconciliation and workflow automation, driven by a fear of falling behind peers.