Prime Ai
Read time: 4 minutes
Leader, welcome back.
$187 billion moved into Anthropic and OpenAI in the last 30 days.
Only 2% of companies have a CFO who owns AI value. In those companies, 3 out of4 say AI is genuinely delivering. Everyone else is running the experiment with no owner!
THE SIGNAL
The CFO seat just changed. Most CFOs haven't noticed yet.
In the last 30 days, Google committed up to $40 billion to Anthropic. Amazon committed $25 billion four days earlier. OpenAI raised $122 billion at an $852 billion valuation and announced plans to spend $600 billion on compute over the next five years.
That capital is going to land on every CFO's desk in 2026 in the form of vendor contracts, token bills, and AI productivity numbers the board will expect to see.
Here's the uncomfortable bit. Babson, MIT, and the Return on AI Institute surveyed 1,006 C-suite executives across 11 countries. Only 2% of organisations have a CFO who owns AI value. Where the CFO does own it, three out of four report AI is genuinely delivering. Everyone else is running the experiment with no owner.
If you're not in the CFO seat, this still matters. The named-owner gap shows up in every function. Wherever AI value is everyone's responsibility, it ends up nobody's.
What does the 2% actually do differently?
I'll spare you the case studies (David Kennedy at Dell ran agents in production for live reconciliations; Ailbhe Moynihan at Meta cut invoice manual intervention from 100% to 7% in seven days; Zachary Wasserman at Huntington runs 50 agents in production with 60 more in development).
The pattern across all three is the same.
The CFO put their own name on the outcome. They picked one workflow to rebuild end to end. They got something into production, not just a pilot. And they tracked a specific before and after metric.
Run yourself through five questions to find out where you sit:
Is there a named owner for AI value in your finance function, not IT, not "the AI committee"?
Is at least one finance workflow running an AI agent in production, not a pilot?
Does your 2026 plan include a specific productivity number tied to AI deployment?
Has your finance team rebuilt at least one workflow around AI in the last 90 days?
Can you name the pound value AI delivered to your P&L last quarter?
If you answered no to three or more, you're in the 98%. The cost of that position is going up every quarter.
Here's how to start.
STEAL THIS PROMPT
The audit prompt that finds your first workflow
Before you rebuild anything with AI, you need to know which of your team's recurring tasks are even candidates.
Run this against your finance team's weekly admin and you'll have a shortlist in 90 seconds.
You are a chief of staff who specialises in process optimisation. I'll paste a list of recurring finance tasks below. Your job:
Group them into logical batches that can be done together.
Identify which tasks can be templated or standardised.
Suggest what can be automated immediately.
Recommend an optimised execution sequence to reduce context switching.
Estimate where time can be saved.
Keep it practical and execution-focused.
Tasks: [paste your team's recurring tasks]
Why this works:
It has a Role (chief of staff), a clear Action (5 numbered tasks), and a Format (practical, execution-focused). That's three of the five elements of the RACEF framework I teach. The two missing ones, Context and Examples, are where you make it 10x sharper. Add a sentence about your team's size, sector, and constraints. Add one example of a task you already automated.
How to use the output:
Take the "automated immediately" list. Pick the task with the highest volume and lowest judgment. That's your first workflow. The 2% started with one. So can you.
The full RACEF playbook, with the 64-prompt library and the prompt chaining workflows that turn 3-hour analyses into 4 sequential prompts, is inside my AI for Finance Leaders course. Currently at a discounted rate for a limited time.
Check it out here.
SIGNAL / NOISE
ChatGPT just landed inside Google Sheets as an extension.
Open a sheet, open ChatGPT in the side panel, prompt it. It builds, audits, and updates spreadsheets in place. If your team lives in spreadsheets, this changes how you draft a model from a blank tab.
Microsoft and OpenAI ended their exclusivity deal on Monday. Less than 24 hours later, AWS announced GPT and Codex are coming to Bedrock. If your firm runs on AWS and you've been waiting on GPT, you no longer have to. Your Azure renewal also just got more interesting.
OpenAI Workspace Agents went live for Business, Enterprise, and Edu users. Codex-powered agents that live inside ChatGPT with a persona, a task, and access to tools like Linear and Slack. They will replace Custom GPTs over time. This is the closest OpenAI has come to shipping the assistant most companies actually want.
The capital is moving in tens of billions a month. The CFOs at the AI labs themselves are publicly modelling restraint. And the data says the only companies actually capturing AI value are the ones where the CFO owns the outcome.
Here's the small move this week. Pick one recurring task on your team's list and ask whether it should still be done by a human in 2026. If the answer is no, that's your first workflow.
Reply and tell me which one you picked. I read every reply.
And if you don't have an AI owner yet and want a second opinion on where your finance team should start, reply anyway. That's exactly what I help finance teams figure out.
