The most challenging part of every AI News of the Month edition is always deciding what to leave out. October was no exception: another month of announcements, launches, and rulebooks.

The main theme this month was accountability.

Not just regulatory, but operational: who owns the output, the advice, and the error when AI becomes part of daily financial work.

Here’s what’s worth your attention—and how to translate it into finance decisions.

The Finance & Accounting Technology Expo is happening in NYC next week. I’m not presenting this time, but I’ll be there — if you’re attending too, feel free to reach out via email or LinkedIn.

Later this month, the AI Finance Club is hosting two sessions:

Both take place just before Thanksgiving, and I hope to see many of you there.

Not a member yet? This is a great time to join the club and see how other CFOs are applying AI to real workflows.

Deloitte Australia Refunds Client for AI-Generated Report Errors

After an AI-written report with fabricated citations surfaced, Deloitte Australia refunded part of a A$440k government contract.

Judicial AI Misuse — U.S. Federal Courts

Two U.S. judges reported rulings drafted with unauthorized AI tools, leading to factual errors.

So what: To me, this was clearly a review failure and control gap in both cases. US Court responded to the issue by “creating a written AI policy and enhancing its review process.” I hope your company has both the policy and the review process!

California is getting its “AI Act” together

California has taken further steps toward comprehensive AI regulation. Governor Newsom signed the AI Transparency Act of 2025 and Senator Wiener’s SB 53, introducing safety standards for powerful AI systems, whistleblower protections, and stricter data-privacy provisions. Other measures address child protection and misinformation.

So what: California’s package is the closest U.S. equivalent to the EU AI Act so far. If you operate, sell, or hire in California, keep track of these regulations — they are real. The chances that this will expand to the federal level remain small, but other states are already taking steps toward their own frameworks.

EU Launched AI Act Implementation Tools

The European Commission rolled out new resources to support AI Act implementation: the Single Information Platform, a central hub with guidance, FAQs, and country materials; the Compliance Checker to help organizations identify obligations; the AI Act Explorer for easy navigation; and a Service Desk staffed by AI Office experts.

So what: Europe is building infrastructure for AI governance. These tools are useful even for non-European companies to test exposure and prepare compliance plans before the Act takes effect.

Claude now allows users to save reusable workflows as Skills. This is a step beyond Claude or ChatGPT projects — a more flexible way to automate repeated processes. Unlike projects that cover a single workflow, a Skill can be reused whenever a similar task occurs, like building presentations, running analyses, or generating reports.

Another exciting development: Claude is coming to Excel. There’s a growing race to integrate AI with spreadsheets — Microsoft Copilot went first, and now Anthropic is following. I joined the waitlist!

So what: The complexity of LLMs is increasing, and so is the need to keep learning. I spent a few hours this weekend experimenting with Skills — it looks promising. I’ll cover it, among other Claude features, in the next AI Finance Club masterclass.

Microsoft added dedicated finance agents for reconciliations and variance analysis to the family of of agents for Finance.

So what: The first release of Microsoft’s AI agents didn’t impress me much, but updates are coming fast and cover more traditional finance workflows. If your team relies heavily on the Microsoft stack, take a look at these agents, they’re starting to deliver real value.

Oracle Fusion AI Agents

Oracle launched integrated agents across ERP, HR, and supply chain modules. Oracle AI agents run on Oracle Cloud Infrastructure and are natively integrated within Oracle Fusion Applications.

So what: “AI-enabled ERP” has been a buzzword for over a year, but now it’s finally visible in real deployments. The capabilities are improving, but user adoption is still uneven.

OpenAI’s new “Atlas” combines browsing, search, and content generation. Anthropic’s Comet appeared earlier, and I had it for a while, but switching from Chrome is still inconvenient, and running two browsers isn’t practical.

The bigger issue: AI browsers are not yet safe. Don’t trust them with company credentials, sensitive data, or passwords.

So what: An AI browser with agentic capabilities is an interesting concept, but until security is addressed, it’s a no-go for business use.

Two important long reads this month:

Indeed Hiring Lab — AI at Work 2025: How GenAI Is Rewiring the DNA of Jobs
Analyzed around 2,900 job skills; 46 percent are already in “hybrid transformation” where AI executes and humans oversee, and only 1 percent are fully automatable.

Wharton – GBK Global AI Adoption Report 2025
82% of leaders use Gen AI weekly and three in four report positive ROI, signaling a shift from hype to measurable value. Data quality, ownership clarity, and implementation discipline make the difference.

There is a lot of conversations on LinkedIn around these reports, but I suggest you read them yourself and make your own conclusions.

I cover my perspective on them in the subscriber-only section

We’ve covered the “what” and the “so what.”

Now, in the subscriber-only section, we get to the real question: what should you actually do with all this?

Not every AI headline requires action — but a few this month clearly do.

The paid edition breaks down the practical next steps: how to tighten review controls after the Deloitte case, what to update in your client contracts and employee policies, and how to budget for AI enablement in 2026.

We’ll also dig into what the latest research from Indeed and Wharton means for CFOs — why AI isn’t taking jobs but reshaping them, and how organizational readiness is now the real driver of ROI.

I realized only recently that it’s been over a year since I started publishing Balanced AI Insights — and just a month since launching the paid section.

Thank you to everyone who’s been reading, sharing, and especially to paid subscribers for your support and thoughtful feedback.

Your messages and insights genuinely shape how this newsletter evolves!

This newsletter is for you, and we want to make it as valuable as possible. Please reply to this email with your questions, comments, or topics you'd like to see covered in future issues. Your input shapes our content!

Want to dive deeper into balanced AI adoption for your finance team? Or do you want to hire an AI-powered CFO? Book a consultation!

Did you find this newsletter helpful? Forward it to a colleague who might benefit!

Until next Tuesday, keep balancing!

Anna Tiomina
AI-Powered CFO