Hi Reader ,

Most finance leaders have tried AI. Fewer have actually built anything with it.

That gap isn't about intelligence or effort, it's about not knowing what's actually possible, and not knowing where to start.

Below are five things that will change how you (and your team) think about AI in finance, a free 60-second tool to figure out exactly what to learn next, and the community to join to when CFOs and Finance Leaders want to stop guessing about AI: the AI Finance Club.

1. AI can build your dashboards — not just analyze the data in them

Most finance professionals use AI to summarize a spreadsheet and stop there. But modern AI tools can go a step further and actually build the interactive dashboard or visual for you — a board-ready KPI dashboard, a waterfall chart, a variance heatmap — as a working file you can open in a browser, no analyst or BI license required.

Try this: Paste a sample of your monthly numbers into Claude or ChatGPT and ask for "an interactive HTML dashboard with a variance chart, a KPI summary, and a filter by department." You'll get something you can hand to your team the same afternoon.

You can even make this dashboard better by asking Claude (this last bit only works in Claude so far) to add an interactive chatbot within the dashboard that can answer questions. It looks like this:

This is one of the first wins people report inside the AI Finance Club , one member, just two weeks into the club, took a sample SaaS dataset and turned it into a working dynamic dashboard, iterating on it himself after learning the basics from the group.

If you want to test the dashboard, respond to this email and I can send you the link for you to test it live!

2. AI can write the Python code that automates your busywork

You don't need to be a developer to use code. If you're tired of manually reformatting the same report, reconciling the same tabs, or cleaning the same messy export every month — AI can write the Python script that does it for you, and explain every line in plain English along the way.

Try this: Describe the manual task step by step ("every month I copy column B into a new tab, remove blanks, and sum by region") and ask AI to turn it into a script you can reuse. Start with a synthetic version of your data, not the real thing (more on that below).

You can then run this code in Google Colab or Microsoft Visual Studio Code:

Members inside the AI Finance Club take this further — one built a fully automated backup system for their finance automation scenarios (daily backups, version control, cost tracking at ~$4/month, all documented in Notion) after a near-miss where a critical workflow was accidentally deleted. Another connected their CRM straight to Claude to auto-generate commissions calculations, payroll notifications, and reconciliation schedules — a process that now runs untouched every month. That's the level the Intermediate and Advanced tracks are built to get you to.

3. You can teach Claude your team's exact workflows — permanently

Most people don't realize you can save a repeatable process — your variance commentary format, your board pack structure, your forecasting method — as a reusable Claude Skill. Once it's built, you're not re-explaining your process every time; you just point Claude at the new data and it follows your playbook automatically.

This is the difference between "using AI" and building a system your whole team can rely on.

More about Claude Skills here:

4. Picking the right model matters more than most people think

Not all AI models are built for the same job, and using the wrong one wastes time and gives you worse answers.

• In Claude: Sonnet is the reliable daily workhorse for most finance tasks. Opus is the deeper reasoning model — reach for it on complex, high-stakes analysis where accuracy matters most. Fable is Claude's newest top-tier model for the most demanding work.
• In ChatGPT: "Instant" is fast and fine for quick questions and first drafts. "Thinking" mode spends more time reasoning step by step — use it for anything multi-step, numbers-heavy, or accuracy-sensitive, like a forecast or a variance bridge.

The rule of thumb: if the task is quick and low-stakes, use the fast model. If getting it wrong is expensive, switch to the reasoning model — every time.

5. If AI couldn't do something last month, try it again this month

AI capability moves faster than almost anything finance professionals have had to keep up with before. A task that failed or felt clunky 30 days ago may work smoothly today — new models, new features, and new tools ship constantly.

The habit to build: don't write off AI for a task based on one failed attempt months ago. Revisit it. The finance professionals pulling ahead right now aren't smarter — they're just re-testing more often.

This is also exactly why we don't think a one-time course cuts it anymore. AI in finance moves fast enough that "I learned it in 2024" is already out of date, you need somewhere that keeps updating with you.

Want a shortcut? Learn inside a community built for finance

The AI Finance Club is a community made for CFOs, controllers, managers, and every finance role in between, with one goal: implement AI in finance in 90 days, without the guesswork.

Instead of chasing scattered YouTube videos and outdated blog posts, members get:

• Live monthly masterclasses — hands-on finance automation, tool strategies, ROI breakdowns, and full replays if you can't make it live
• An AI Finance Expert Network — get answers in 24 hours and learn from peer examples and reviews, not just theory
• A monthly "State of AI in Finance" briefing — the tool launches that actually matter, security & compliance updates, and actionable summaries
• Process deep dives — step-by-step playbooks with the exact tools and prompts used, built to save tens of hours a month

Choose your learning path

There's a clear track depending on where you're starting from:

So — what should you actually learn next?

That depends on your role, your current AI comfort level, your tools, and your biggest pain point. Guessing wastes time; a 60-second diagnostic doesn't.

I also built a free calculator that asks six quick questions and gives you back:

• Your recommended AI tool (Claude, ChatGPT, Copilot, or Gemini — based on what you actually have access to)
• An AI Learning Priority Score
• Your next 5 topics to learn, in order
• A practice project matched to your biggest pain point
• A 30-day learning plan you can actually follow

👉 Take the "What Should You Learn About AI?" ​

It takes less time than reading this email.

Think of it as step one. Step two, if you want a structured 90-day path instead of figuring it out solo — with live masterclasses, an expert network, and monthly updates so you're never behind — is the AI Finance Club.

Remember, the finance teams pulling ahead are not waiting for AI to become perfect. They are testing, building, and learning faster than everyone else.

Start with the free diagnostic, find your next move, and then go build something useful.

Thanks for reading and hope this helps!

Christian Martinez