Read time: 3 minutes

I've been in rooms with founders, CFOs, and operators who all say the same thing - "I feel like I'm falling behind."

And I understand it.

When your feed is full of people building and shipping, it's easy to feel like the train has left.

But here's the thing.
That feeling comes from the bubble you're in, not from reality.

Let me show you the chart that changed how I think about this.

84% of the world has never had a single conversation with AI.

Let that land for a second.

Every dot on this chart represents 3.2 million people.
The entire grid is 8.1 billion humans.
The green strip at the bottom? That's the 16% who've tried a free chatbot.
The yellow sliver above it? The 0.3% paying $20 a month.
The barely-visible red line at the top? The 0.04% using tools like Cursor or Claude Code.

If you're reading this, you're somewhere in that yellow or red sliver.

Which means your whole sense of how "mainstream" AI is comes from a bubble so small it's almost invisible on the chart.

I get it.
I've felt it too.
Surrounded by builders and creators - online and in real life - everyone seems to be shipping something. It's easy to feel like you're already behind.
But that feeling is the bubble talking, not reality.

Think about where the internet was in 2005.
Only 16% of the world was online.
Everything we take for granted today - social media, e-commerce, streaming, cloud, SaaS - was built after that point.
The entire wave came later.

That's where AI is right now.

The companies that move AI from 0.3% to the next billion users will define the next decade.
The professionals who show up early, learn the tools, and build the skills now will own the space when it goes mainstream.
That shift will happen faster than the internet did - and the window to get ahead is still open.

So keep reading. Keep learning. Keep applying.

That's exactly what we're doing here - and I want to know where you're at with it.

Hit reply and tell me: what's the one AI workflow you're actually using in your work right now? I read every reply.

A London-based startup called Stacks just raised $23 million in February to automate finance operations with AI agents.
Variance analysis, journal entries, reconciliations, month-end close - all automated. They've already saved finance teams across 30 enterprise customers over 100,000 hours annually since coming out of stealth less than a year ago.

And they're not alone.

Deloitte's latest CFO Signals Survey found that 54% of finance chiefs are now making AI agents a digital transformation priority this year.

A Fortune 500 CFO who spoke to CFO Dive this month said her team has already cut their financial reporting cycle by 40% using an internal agentic AI tool built with Deloitte.

My take: the Big 4 deploying AI agents in production, startups rebuilding the CFO's office from scratch, and over half of finance chiefs now treating this as a 2026 priority - this is no longer a "wait and see" conversation.
The question for finance leaders isn't whether to adopt. It's how fast.

A new study by The Futurum Group, where 830 IT decision-makers were surveyed in early 2026 - found something significant:
"Productivity gains" collapsed as the #1 ROI metric for AI, falling nearly 6 points in a year.

In its place, direct financial impact - revenue growth and bottom-line profitability - nearly doubled to become the new primary measure of success.

Simultaneously, agentic AI surged 31.5% year-over-year as a top technology priority.
The pilot phase is over.
Enterprises are now asking: what did this actually do to our numbers?

My take: this is the right question to be asking.
AI that can't connect to business outcomes is just an expensive experiment.
The finance leaders who frame every AI initiative around P&L impact - not just time saved - will be the ones who get investment, buy-in, and results.
If you're building a business case for AI at your company, this framing matterS

Before we touch anything else, we build one skill together.

In Claude, a Skill is a saved workflow - a set of instructions you build once and run on repeat. Instead of typing out a long prompt every time, you open the skill and it already knows exactly what to do

Not a strategy.
Not a roadmap.
One skill - a saved, repeatable workflow for a task they do constantly and hate spending time on.

Here's exactly how I walk them through it.

Step 1 - Pick one task you do repeatedly:
Writing emails, drafting reports, creating briefs.
Something you do often enough that automating it actually matters.
Don't start with your most complex task.
Start with the one that wastes the most of your time for the least strategic value.

Step 2 - Write down how you'd explain it to a smart intern:
What context do they need to do it well?
What does good output actually look like?
What mistakes do you always end up correcting?
Write that down. That's your prompt foundation.

Step 3 - Add examples:
Show two or three inputs and what the ideal output looks like for each.
This is the step most people skip and it's the one that matters most.
Examples beat instructions every time.
The model mirrors what you show it far more reliably than what you tell it.

Step 4 - Include your voice:
Paste in samples of your own writing - emails, commentary, reports you're proud of. This is what makes the output sound like you instead of generic AI.
Without it you're just getting the model's default style, which sounds like everyone else using the same model.

Step 5 - Save it as a skill:
Now you're not prompting from scratch every time.
You open Claude, run the skill, done.

Start with one.
Get it working.
Then build the next.

Within a month most of my clients have five or six skills running.
The cumulative time saving is significant - but the bigger shift is that they stop thinking about AI as a tool they use and start thinking about it as a workflow that runs.

If you want help building your first skill for your finance team, hit reply. It's something I work through with clients directly.