Tina Feuerstein made more than 3,800 personal purchases and took about $1.09 million from the company, buying everything from Coach and Michael Kors to furniture, family expenses, groceries, and utility bills.

A jury convicted her on eight counts of wire fraud. Each count carries up to 20 years.

Here's the part that should bother finance people...

She was not some random employee gaming the system.

She was the CFO.

She also had access to the accounting system and, prosecutors said, used it to hide the spending.

So...

A lot of fraud controls are built around the idea that the CFO is the one enforcing them.

But what happens when the CFO is the one walking around them?

What’s the better protection against executive fraud?

A) tighter board-level review of spending

B) stronger system limits, even for top finance people

C) independent internal audit with real teeth

An AI company once worth $1.4 billion just got hit with the most embarrassing fraud story possible:

According to prosecutors, the AI was not the fake part.

The revenue was..

The former CEO and CFO of iLearningEngines were charged after prosecutors alleged they fabricated “virtually all” of the company’s customer relationships and revenues,

inflating sales by hundreds of millions of dollars a year and, at times, by more than 90% of annual revenue.

And the details are ugly:

The point is..

When markets get drunk on a theme, in this case AI, weak numbers can hide longer because nobody wants to be the person asking boring questions while everyone else is chasing the next unicorn.

That’s how nonsense gets room to breathe.

Question for finance pros: What makes fraud easier to sell in a hot market?

You can understand cash flow, accruals, working capital, margins, and forecasting perfectly.

But the minute you explain them in technical language, many founders hear:

“Your numbers are complicated, and this conversation is going to be painful.”

That’s where the trouble starts…

I recently came across Hillel Steiner CPA, MBA, a fractional CFO firm owner and author of Smart Numbers, Real Growth.

The book was written for business owners without a finance background.

And that’s why it’s useful for CFOs, CPAs, and bookkeepers too...

Because it gives you something most finance professionals need more of:

Founder-friendly language.

Inside, Hillel explains (in founder-language) ideas like:

This is the kind of book you can recommend to founder clients before they work with you.

Or use it as a reference for how to explain financial concepts in language they actually understand.

Check out the book here

Prefer the free PDF version? DM Hillel here

That’s a rough reaction for a company trying to sell the market on “Fermi 2.0.”

The company says the shake-up is part of its move from startup mode into a more professional, public-company setup. Fine.

But investors usually hear something else when:

Especially when the business is tied to one of the hottest and most capital-hungry themes out there:

building power infrastructure for AI.

And that’s the real tension here.

Fermi is selling a huge future: private electric grids, massive power demand, AI-scale infrastructure.

But the market still wants the boring basics:

stable leadership,

clean governance,

and someone who looks firmly in charge...

PwC says 81% of C-suite leaders think their companies are still at least a year away from seeing meaningful AI returns beyond basic efficiency gains.

And yet most still plan to keep spending.

That’s the tension finance people should care about:

The market is basically saying:

PwC’s Dan Priest says the problem is not that AI can’t work.

It’s that most companies are stuck between pilot and real transformation.

They’ve tested pockets of value, but haven’t redesigned the workflows, roles, and decision-making needed to get bigger returns.

So now CFOs get the fun job:

keep funding AI while also explaining why the payoff still looks like a future promise

Cut 16% of staff, talk up AI, then hand the CFO seat to the internal finance operator closest to the model.

Doug Hott is taking over as CFO after Derek Anderson, CPA’s exit, right as Snap is..

laying off about 1,000 employees, closing 300 open roles, and telling investors AI will help do more with less.

The company says the cuts should lower its annualized cost base by $500M+ in the second half of the year.

Snap is under pressure to get leaner, catch up in AI, and look more serious about profitability, all at the same time.

So the next CFO’s job is pretty obvious:

And yes, the activist letter in the background makes the whole thing louder.

keep saying they want AI ROI.

Then they approve pilots with no scoreboard.

A new press piece says the fix is painfully simple:

pick the exact KPI first. Not “transform finance.” Something real like:

That’s what West Monroe’s Connor Augustyn is arguing after hearing the same complaint from CFOs over and over:

And honestly, that’s the whole problem.

Too many AI projects got approved because:

So a year later, finance is left asking the awkward question: what exactly did we buy?

Augustyn’s advice is to get specific, and start with the boring stuff that already works,

like invoicing and reconciliations.