Why Finance Leaders Should Use AI for Improved Efficiency
AI is already improving how companies work. Finance is lagging, but not for long. As Axel Demazy, CEO of Spendesk, puts it: "Gen AI has accelerated access to AI; and when you look at how steep that acceleration is, you can tell the inflection point is far from being reached. We have only scratched the surface of how AI can augment human operations at scale in all our workflows, and we will see continued investment by businesses as a result."
Spendesk's report, The State of AI in Finance, is based on hundreds of conversations with CFOs across the CFO Connect community. The takeaway: efficiency gains are real, and a growing group of finance leaders are proving it in the field.
Adoption is moving-but most teams still haven't started
Finance teams made progress in the past year, yet 61% still haven't implemented AI in their workflows. At the same time, 85% of CFOs are optimistic about the efficiency upside.
What's holding them back? Mainly ROI clarity, skills, and time.
- Unclear benefits of AI (36%)
- Limited skills or training (32%)
- Time constraints (14%)
- Data privacy or security concerns (12%)
- Lack of leadership support (5%)
Many leaders still see AI as a helper for emails or simple questions-useful, but not core. Others worry their teams aren't ready. As Finup360's Founder Anne-Claire Chanvin says: "This shift is as big as when Excel was introduced in the 80s."
Where AI delivers value right now
AI handles rule-based, repetitive, time-consuming work well. That frees your team to focus on judgment, scenario thinking, and strategic callouts.
Sarah Fu, Founder and Managing Partner of Elsa Capital, highlights the obvious targets: weekly and monthly refreshes, standard slide decks, and model rebuilds that eat hours. These are predictable processes, which makes them ideal for automation with LLMs and scripts.
Bogdan NΔforniΕ£Δ, CEO and Co-Founder of Profluo and former BMW CFO, sees the human side: "We keep seeing demotivated junior staff stuck doing basic work. No wonder there's such high turnover in finance teams." Better tooling helps retain and upskill talent.
High-impact use cases you can deploy this quarter
- Reporting refresh and variance commentary: Automate data pulls, draft variance explanations, and pre-build slides. Anne-Claire Chanvin shared a case where a task that "previously took one day per accountant, per country" took "five minutes" using a ChatGPT script.
- Invoice and expense processing: Extract line items, auto-code to GL, flag policy exceptions, and summarize vendor-level trends for review.
- Policy and memo consistency checks (ASC 606): Paul Jun, former VP of Corporate Finance and Strategy at Dropbox, notes that an AI tool can compare your ASC 606 memo against the handbook and model contracts for consistency-like a second set of eyes. For context on the standard, see the AICPA's overview of ASC 606 here.
- Document reviews and internal quality control: Use AI to check accounting memos, investor updates, and board materials for clarity, policy alignment, and numerical consistency before they reach auditors or executives.
- Email and presentation drafting: Standardize tone, formatting, and terminology across teams. Faster drafts, fewer edits.
- Month-end close assistance: Auto-match transactions, propose reconciliations, and surface anomalies for controller review.
Why adoption stalls-and how to move
A late-2024 CFO Connect survey showed most leaders see the upside but don't know where to start. Here's a simple plan that works without derailing BAU.
- Pick 2 high-friction processes: Think monthly reporting refresh, invoice coding, or expense policy checks. Time them now to set a baseline.
- Set guardrails: Define data access rules, PII handling, approval steps, and where human sign-off is required.
- Start with proven tools: Use vendor solutions or secured LLMs for one pilot, not ten. Keep scope tight and measurable.
- Train for 60-90 minutes, then practice: Short live sessions, simple playbooks, and a shared prompt library. If you need a curated starting point for tools, see AI tools for finance here.
- Measure ROI weekly: Track hours saved, error rates, rework, and cycle times. If it's not saving at least 30-50% time after month one, adjust or stop.
Privacy and control without slowing to a crawl
- Use enterprise plans with SSO, audit logs, and retention controls.
- Redact sensitive data and restrict uploads of PII, payroll, or legal exposure items.
- Keep humans in the loop for journal entries, policy interpretations, and revenue recognition decisions.
- Standardize prompts and workflows; maintain version control like you do with models and policies.
The bottom line
Finance leaders aren't late-they're early enough to still choose their edge. The Spendesk report shows hesitation is mostly about clarity and capability, not a lack of potential.
Start small, ship one meaningful use case, and let the numbers guide your next step. The teams that do this now will reclaim hours, reduce churn, and raise the bar on control and decision quality-without waiting for a perfect playbook.
Your membership also unlocks:
