Today in Finance AI
- 1Global regulators just moved in sync on AI oversight.
A September regulatory roundup covers ASIC naming AI a 2026-27 corporate priority, watching deepfakes, AI-driven customer decisions, and market integrity, while the EBA, EIOPA, and ESMA have jointly called for coordinated supervision of AI-related ICT risk, consultation open until Dec 31, public hearing Sept 29.
- 2Token usage is starting to show up as a revenue signal, not just a cost line.
A BCG analysis found companies in the top quintile of AI token usage saw 16.5% median year-over-year revenue growth, against 5.1% for the lowest-usage group — real evidence heavier AI usage is starting to correlate with business outcomes, not just spend.
- 3The automation ceiling in finance is real, and it's exactly what today's tool comparison is about.
Most AR automation software gets a team to 60-80% touchless processing, then stalls — the last 20-40% sits in an exception queue because no two customers pay the same way: different portals, different paperwork, different person to chase.
An unexpected question on margin impact shouldn't derail an FP&A team's weekend or produce outdated Monday reports. Learn how predictive analytics automates data unification and sensitivity forecasting, equipping CFOs to respond to executive scenario requests with actionable, decision-ready data in under an hour.
System Connectors & Integrations: Pick by Problem, Not by Feature List
Mid-market, ERP-centric (especially NetSuite), operations-focused, two-way sync. $600–$6,000/mo depending on transaction volume — the lowest-cost entry point of the three.
One-directional only, moves ERP data into a warehouse (Snowflake, BigQuery) for reporting. Doesn't write back, ruled out if you need transactional sync. Consumption-based, SMB spend averages ~$30K/yr, enterprise ~$110K/yr.
Enterprise-grade workflow automation across many apps at once, not just two-system sync. 1,200+ connectors, usage-based pricing, priced well above Celigo, built for automating a whole business process, not moving data.
Where would you most want AI to help your finance function?
That's it for this week.
Keep executing, not just experimenting. See you next week.
The CFO AI Hub Team