Three weeks ago, Larry and I started this series with an uncomfortable number: 59% of finance functions claim to be "using AI," but only 11% have meaningfully implemented it. 

In Edition 1, we diagnosed the problem. In Edition 2, we introduced the solution: the Finance AI Center of Excellence (CoE), and many of you took the Red Flags Diagnostic to see where your organization stands. 

This week, we close the loop. The question I hear most from readers and clients is not "should we build a CoE?" It is "how do we actually start?” Larry argues that the real differentiator is not technology, but orchestration. The answer increasingly points to a formal Finance AI Center of Excellence (CoE) that institutionalizes the governance architecture and resource utilization that transforms AI from isolated tools and complexity into sustained capability and competitive advantages. 

Larry will walk through the structural blueprint: design principles that work as an interconnected system and the operating model that holds it together.

I will add the practitioner layer: what to build first, how to staff it with a small team, and a 90-day roadmap you can start immediately. 

HAPPENING TODAY: Larry and I are presenting "Stages of AI Maturity in Finance" at the Argyle AI Solutions Expo. It’s not too late to join!

Thursday, Feb 26th: Building with Claude – Excel, MCP Connections & Cross-Tool Automation. This is the third Claude Masterclass I'm hosting with AI Finance Club. We'll get hands-on with Claude in Excel, the Chrome extension, MCP connections (Stripe, PowerPoint, Word, and how to set up your own), plus the newly released Claude Cowork.

If you're already part of the AI Finance Club, don't miss this one. New to the club? Nicolas Boucher runs a free masterclass that gives you a look inside before you join.

A Finance AI CoE is not an IT support group. It is a governance and integration layer embedded within the Office of the CFO. Its mandate includes: 

In short, the CoE institutionalizes AI-enabled Decision  intelligence. 

In Edition 2, we walked through the value propositions of a Finance AI CoE. Larry takes it further in this final installment: the design principles are not a menu to pick from. They are an interconnected system where weakness in one undermines all the others. 

The five design principles:

Finance Ownership. AI in finance cannot be delegated entirely to IT or analytics teams. The CFO must retain accountability for outcomes, risks, and controls. Finance ownership ensures AI aligns with financial governance, regulatory expectations, and enterprise priorities. 

Embedded AI. CFOs consistently prefer AI embedded within core finance platforms (ERP, EPM, close, and spend systems) rather than standalone tools. Embedded AI strengthens adoption, reduces integration risk, and prevents the fragmentation that comes from parallel processes producing inconsistent outputs. 

Controls by Design. Finance cannot afford "black box" AI. Explainability, auditability, and model governance must be designed from the outset, not retrofitted after something goes wrong. 

Value Measurement. AI initiatives must be evaluated like any other finance investment, tied to KPIs that finance already tracks days to close, forecast accuracy, cost per transaction, exception rates. 

Talent and Roles. AI does not eliminate finance roles; it reshapes them. A Finance AI CoE provides a structured approach to role redesign, upskilling, and capability development—moving finance professionals toward higher-value analytical and judgment-based work. 

Why does the system matter more than the parts? These principles are interdependent, and that is the critical insight. Finance ownership without controls invites risk. Embedded AI without value measurement erodes credibility. Talent development without governance leads to fragmentation. You cannot build one pillar and ignore the others. The CoE exists to coordinate and reinforce all five simultaneously. 

Anna's take: For smaller teams, the interdependency point is both the challenge and the advantage. You do not have the luxury of building each pillar with a dedicated team. But you also do not have the bureaucracy that slows down large organizations. The non-negotiable starting point is finance ownership. If the CFO does not own this, it drifts to IT and becomes a technology project instead of a finance capability. Everything else is built from that. 

Larry outlines four layers that a functioning CoE integrates: 

1. Model and Analytics Layer. Where predictive, generative, and agentic models are developed, validated, and controlled. Think: your forecasting models, your AI-generated commentary, your anomaly detection tools. 

2. Governance and Risk Layer. Defines validation standards, bias controls, data lineage requirements, access protocols, and audit documentation. This is the layer that answers the question "can we trust this output?" 

3. Operational Integration Layer. Embeds AI output into the processes that matter forecast cycles, dashboards, board reporting, capital allocation decisions. Without this layer, AI stays in the lab. 

4. Executive Oversight Layer. Aligns AI initiatives with strategy, value creation priorities, and risk appetite. This is where the CFO connects AI capability to business outcomes. 

When these layers function together, AI becomes systemic and embedded rather than disjointed and experimental. 

Anna's take: If you are a small or mid-size finance team looking at these four layers and thinking "that is a lot of infrastructure," here is the practical version. Start with layers 2 and 3. Governance and operational integration. Get your controls right and embed AI into one or two real workflows. Layers 1 and 4 will develop naturally as you scale. You do not need all four layers on day one.  

One more reason this matters now: AI in finance is moving from tools you use to agents that act. 

As finance approaches agentic AI, systems capable of triggering workflows, adjusting forecasts, or flagging exceptions within defined guardrails, governance becomes even more critical. Autonomy without governance introduces risk. Autonomy with orchestration creates competitive advantage. 

A CoE ensures that agentic systems operate within defined thresholds, escalate exceptions appropriately, preserve auditability, and maintain executive oversight. 

Anna's take: This is not a future concern. It is already happening. If your finance team uses AI to draft variance commentary, run first-pass reconciliations, or generate scenario analyses, you are already managing AI agents in some form. The question is whether you are doing it deliberately or accidentally. In Edition 2, we talked about managing mixed teams of people and AI agents as a new competency. The CoE is where that competence gets built and governed. 

What Not to Do 

The Leadership Imperative 

AI will not replace finance leaders. But finance leaders who understand AI orchestration will redefine the function. The future of finance will not be defined by how many AI tools an organization deploys. It will be defined by how well those tools are governed, integrated, and aligned with strategy. That is the role of the Finance AI Center of Excellence.

That is the full blueprint: the principles, the operating model, and the direction AI is heading.

Now the question is where to start and what to do in the first 90 days. In the subscriber section, I break it down into a phased roadmap you can start on Monday, plus role definitions that work evenfor small finance teams.

That wraps our three-part series with Lawrence Maisel.

We started with the uncomfortable gap between AI activity and AI capability. We moved to the structural solution and the warning signs to watch for. And now you have the blueprint and the roadmap to start building.

I want to thank Larry for collaborating on this series. Bringing together his enterprise-level expertise and my startup and mid-market experience has been exactly the kind of cross-pollination this space needs. The problems are the same regardless of company size.

If you're thinking about establishing an AI Center of Excellence but aren't sure where to start, or you need help driving the initiative from strategy to execution, that's exactly what I do. Book a call or respond to this email, and let's talk about what makes sense for your team.

And if you want to reach out directly to Larry, here’s his email [email protected].

We are working on packaging the complete series, including all three editions, the diagnostic tool, and the implementation materials, into a comprehensive guide. Stay tuned!

In the meantime, I would love to hear from you. Are you planning to build a CoE? Already started? Think it is overkill for your team? Hit reply. I read every response.

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