Read time: 5 minutes

Leader, welcome back.

Nobody knows their perfect first AI workflow before they start.
The ones who figure it out fastest are not the best planners.
They are the ones who notice what they are already doing on repeat, try a few approaches, and go deep on whichever one actually clicks.

Wide first. Then narrow. Then own it.

Here is the workflow most finance leaders are still doing manually.

Research a topic, open PowerPoint, stare at a blank slide, paste in what you found, reformat everything to match the template, write the commentary.
Two to four hours, minimum. Every single time.

Here is the alternative.

Start with a Claude Project.
Write a system instruction that tells Claude who you are, what the business does, what your outputs usually look like, and what good looks like to you.
This takes 20 minutes once. You never have to explain yourself again.

Then do the research inside that project.
Feed it earnings calls, filings, internal documents, messy PDFs.
Ask it to synthesise and surface what matters. The context stays live across the whole session.

Here is where it gets interesting.

As of March 11, Claude shares full context across Excel and PowerPoint simultaneously.
You ask it to pull data from your workbook, build the analysis, and populate your slide deck, all in one continuous conversation. No switching tabs.
No re-explaining the dataset. The deck comes out in your template, your fonts, your layout.

When the workflow lands right, save it as a Skill.
One click next time.
Same output, without the setup.

That is the board pack workflow. That is the competitor benchmarking workflow. That is the investor update workflow. Same pattern, every time.

And once you see it, you see it everywhere.
I know a team running a weekly competitive intelligence report that pulls from earnings calls, and analyst commentary automatically every Friday.
Work that used to take hours of manual pulling and reading, now a digest waiting in their inbox.

I know finance leaders building DCF models and complex analysis from raw spreadsheets and clunky PDFs in hours, work that honestly would have taken me weeks early in my career.
One CoWork setup processes 20 industry emails a day into a five-minute read so you get the signal without the noise.

Another team fed their collective thinking into Claude, built a full digital strategy across multiple brands, loaded it into their project management tool with owners, timelines, and task descriptions, in under a week.
Work that would have cost tens of thousands with a consultant.

The pattern is always the same. You are not doing something new. ~
You are doing what you already do, starting from 80 percent instead of zero.

The only question is which task you are going to start with.

Want to bring AI into your finance function properly?

We work directly with finance leaders and their teams to design and implement AI workflows that save real hours, not just in theory.
From month-end automation to board pack generation.

Hit reply or talk to our team here.

The average month-end close still takes 6.4 business days.
According to AFP, 78 percent of finance teams are still tracking it in spreadsheets.
A joint MIT and Stanford study found that teams adopting AI in their accounting workflows reduced close time by up to 75 percent.

The technology to cut your close by three quarters is available right now, today, and most teams are not using it.

The AFP FP&A Forum ran a session this month with a straightforward title: "How AI Saved 50 Hours in My Monthly Finance Operations."
One CFO. One finance function. Fifty hours recovered from building models, identifying anomalies in accruals, and stress-testing cash flow assumptions.
That is more than a full working week, every single month.

A finance director with thirty years of experience, said something at a CFO Connect session recently that I keep coming back to.
The top priority for finance leaders right now is faster decision-making.
The biggest barrier is not the tools and not the budget. It is fear.
Fear delays adoption, and delayed adoption turns AI into a firehose you are scrambling to catch up with rather than a system you built deliberately.

The CFOs saving 50 hours a month did not start with a transformation roadmap. They started with one painful task and timed how long it took.

What is yours?

Two launches worth watching:
Perplexity and Manus both shipped desktop AI agents this week that live on your hardware and hold memory between sessions.
The detail that matters for finance teams: files stay local, only task instructions reach the cloud. That is the answer to the security objection that has been blocking enterprise AI adoption in environments where data sensitivity is the whole conversation. IT can finally tell legal a story they will actually accept.