News of the Day -

AI is spreading across finance teams, but the cash cycle keeps getting worse. A new Auditoria.AI report finds the average finance team now runs AI across 2.43 functions, up 36% in a single year, yet the share of teams spending 11+ hours a week chasing vendors and customers jumped to 44.5% this year, the largest single-year move the report has tracked. More AI adoption isn't the same as fixing the actual bottleneck. Auditoria.AI
CFOs are warning AI can't clean up a finance department that was already a mess. At a CFO panel this week, executives from Cigna and Northwest Bankshares said AI is only as reliable as the data feeding it, and that speed without trust becomes a liability, not an advantage. Several agreed the real work is fixing data ownership and process discipline first, not adding more tooling on top. Thomson Reuters
A new "always-on" AI CFO just launched, built for founders who don't have a finance team yet. cfo.ai (formerly Runway) launched Ari, which connects to a company's financial systems, builds and maintains a live financial model, forecasts cash, runs scenarios, and proactively flags what needs attention rather than waiting to be asked. Worth watching whether this category moves upmarket toward larger teams next. PR Newswire
Article of the Week -
Why SMBs Have the AI Advantage

While large enterprises burn budget on uncontrolled AI experimentation, so-called "token maxxing," SMBs have a real edge: they can skip the costly trial-and-error entirely. Pairing strict financial guardrails (the GFIS framework) with plug-and-play, outcome-focused AI implementations from GroByz Partners and Selona AI lets smaller finance teams secure ROI immediately instead of funding someone else's learning curve. The advantage isn't resources, it's speed. Read more.
The Shortlist -
Case Studies:
OpenAI built an internal Contract Reader Bot that extracts contract terms and applies ASC 606/IFRS 15 revenue recognition logic automatically, generating journal entries without manual review at first pass. Result: their finance team reportedly runs at roughly 22% of the headcount of comparable tech companies.

Morgan Stanley rolled out Debrief, an AI assistant for its financial advisor teams. Result: 98% adoption across advisor teams, follow-up tasks moving from taking days to taking hours.

Klarna deployed an AI customer service assistant that handled 2.3 million conversations in its first month. Result: resolution time dropped from 11 minutes to under 2, repeat inquiries fell 25%, workload equivalent to roughly 700 full-time agents, an estimated $40M profit improvement in 2024, independently corroborated by Reuters.

Verdict: three different kinds of proof, not three tools competing for the same job. OpenAI shows AI replacing one specific technical judgment call. Morgan Stanley shows adoption at scale inside a regulated, relationship-driven function. Klarna shows what the case looks like measured in company-wide P&L impact rather than hours saved on one task.
Upcoming Events:
From AI Curiosity to Operational AI Mastery — Online webinar, Tuesday 29 September, 11:30am-12:30pm UK time. Hosted by GroByz, Selona AI and Guftagu. A practitioner walks through what's actually running week to week in cash flow forecasting, month-end close and GST/VAT reconciliation. Free to attend. Register
Scale Your Finance Function: AI Playbook for Faster, Accountable Operations — Online webinar, Wednesday 7 October, 11am-12pm EDT. Hosted by Neoflo. Nipun Mehra (Founder & CEO, neoflo.ai) on building a controlled workflow layer across Procure-to-Pay, Order-to-Cash and Record-to-Report without replacing your ERP. Register

That’s it for this week.
Keep executing, not just experimenting. See you next week.
The CFO AI Hub Team
