AI in FP&A: What Changed in 2026, and What a Small Finance Team Should Do

🤖 1. AI now works inside the tools you already use

In June 2026, Microsoft added finance "skills" to Copilot in Excel: repeatable steps for building a DCF, closing the books, refreshing a monthly reporting model or preparing a variance analysis.

Planning platforms are moving the same way. Anaplan launched new AI tools in March 2026, including Agent Studio for building AI agents. Workday and Oracle now offer predictive forecasting and plain-English questions on planning data and reports.

📊 2. Everyone is trying it. Few see the payoff yet.

• 59% of finance teams use AI (Gartner, 183 finance leaders, 2025)
• Only 7% report a high impact (Gartner)
• Only 21% report clear, measurable ROI (Deloitte, Finance Trends 2026)

🧹 3. The bottleneck is data, not AI

61% of FP&A teams say unreliable data holds back their technology, and 96% still plan in spreadsheets (AFP, 2025 FP&A Benchmarking Survey).

AI on messy data gives you wrong answers, faster.

🎯 4. Start small, and keep it reviewable

Pick one monthly task where finance can check every output: variance analysis, a first draft of the commentary, reconciliations or anomaly flags.

AI prepares. Finance approves.

✅ Where to start this quarter

• Pick one process and time how long it takes today
• Fix your definitions first: accounts, SKUs, customers, departments
• Try the AI features in tools you already pay for
• Keep a human sign-off on every output
• Set clear rules for what data can go into AI tools
• Measure for 60 to 90 days, then decide whether to expand

💬 Which part of your month-end would you hand to AI first, and which part would you never hand over?