📬 Your AI product has a fake mustache...

Sick of typing the same instructions into Claude every single time? Skills work differently. You teach Claude a workflow once, and it runs the same way every time. 

Aleph’s Claude skills for finance guide gives you 6 ready-to-rock skill files, including:

You know, the kind of thing you’d build yourself ‘when you get the time’...

👉🏼 Feeling the weight of a complicated problem and need advice? Send me your questions, and you might just see yourself in next week’s Mailbag. Submit anonymously using the button below:

Here’s what’s cooking:

Now, let’s get into it.

AI in Name Only from the US 

My company is in the services industry and is going to market with a new bundled product, which is being masked as an "AI" solution. However, when you dig under the hood, it is an "AI in Name Only" solution. Of the services within the bundle, 80-90% are not necessarily AI-focused. While our systems support visibility into each component of the bundle, the executive team is deciding to lump all revenue under AI - from sales opp, to invoice, to how the revenue hits the GL - limiting the true breakout for reporting purposes. 

Understanding that AI is the latest buzzword, and this is very likely a marketing tactic, how would you approach this situation when there are concerns about authenticity of true revenue composition? 

P.S. Nothing fraudulent, just a difference in classification expectations.

I think this one is fairly simple.

How you market a product and how you classify and report it internally are two different things.

Your company is taking a bunch of existing services, wrapping them in some AI glitter, and trying to use that packaging to drive better volume, margin, or market attention.

Join the club… so is almost every business on the planet right now.

And to be fair, that may not be dumb. If the market wants AI, and you can package a more valuable solution around that theme, fine. That is a commercial decision.

But finance needs to understand what is actually being sold.

So, the question is not “should we market this as AI?” That is for sales, product, and marketing to argue about in whatever room smells most like brand strategy. The finance question is: what is the economic impact of the re-bundling?

Are you getting better pricing? Better gross margin? Better win rates? Better retention? Better expansion? Better acquisition economics? Are customers buying the bundle because of the AI component, or because you stapled existing services together and put a shiny label on the front?

You cannot answer those questions if you lose the component-level data at source.

I would be very clear: the business can choose how it packages the product externally, but finance must preserve visibility internally.

That means tagging the revenue properly at sales opportunity, contract, invoice, and GL level. You can always aggregate later. You can always decide to present “AI revenue” as a bundled commercial category in a board pack or investor story, assuming that is not misleading. But once the underlying data is lost, that optionality is gone.

And I do not understand why the executive team would resist that.

They should want the atomic visibility. If this AI packaging works, they will want to know why. If it does not work, they will want to know that too. If margins improve, they will want to know which component drove it. If churn worsens, they will want to know whether the bundle created confusion. If customers complain that the “AI solution” is mostly just old services wearing a fake mustache, that is useful information.

So yes, I think you should push back. If it is controversial, I would wonder what people are trying not to see.

My advice on this issue is simple: get hold of the classification now. Define the product tags. Preserve the component breakout. Then let the business decide later how to aggregate and present it.

But one other thing for you to think on... this sounds symptomatic of a bigger cultural issue. Finance should not be reacting to how the exec team wants to classify revenue. Finance should be leading the design of how revenue gets captured, tagged, controlled, and reported. So, I’d worry about how little influence you seem to have on this conversation. Why are you being told and not doing the telling? Are there other finance dynamics you are in the passenger seat on when you should be driving? Why is that?

TLDR: Don’t confuse commercial packaging with internal reporting and classification. 

YouAreTheMan from the US

I work for a manufacturing company that uses standard costing. I want to know what a mature standard costing governance/process looks like. For example, do you update assembly scrap % once a year and update labor/OH rates yearly? Do you need to conduct an annual review of your parts to check if any of them should change from active to inactive, or an annual update of your raw material costs? And when you perform a cost roll there should be zero costing errors, etc? 

What else am I missing? I essentially want to know what policy, governance, controls I should put into practice so that standard cost is as accurate as possible.

Thank you for everything you do!

I’m noticing a pattern.

I’m getting an increasing number of questions about the detail of manufacturing and standard costing.

It’s almost like my readers are starting to figure out this is a topic I love talking about. And that if they ask a question about the glorious detail of purchase price variances, standard cost rolls, or overhead absorption, they are almost guaranteed to get featured in the Mailbag.

Well, I’m onto you.

And I regret to inform you that you would be… ABSOLUTELY RIGHT.

It’s like when I’ve been bugging my kids to go to bed for the last hour, and they suddenly say, “Dad, tell me about the time you came off the bench to score an overhead kick on debut for your school team.” (True story)

Or, “Dad, what is your favorite Otis Redding song?”

Or even, “Dad, tell me again what EBITDA stands for.”

They know what they are doing.

Anyway, onto your question.

I would think about standard costing governance in three buckets:

Let’s start with the annual rollover.

The clue is in the name. In a mature business, you generally want to roll standard costs once per year. The whole purpose of standard costing is to give you a fixed reference point for performance, so you can break complex price, volume, mix, yield, usage, labor, and overhead variances down to product level.

If that reference point moves too frequently, it stops being a reference point. It becomes another moving part in the machine.

A movement in standard margin is usually about volume, mix, portfolio, pricing, or structural margin; i.e. something commercial. A variance is usually about purchasing, manufacturing, labor, yield, efficiency, or overhead absorption; i.e. something operational.

You are not just putting dollars into different lines of the P&L. You are pointing accountability at different parts of the business.

So, the annual roll needs to be robust.

That means a proper timetable, clear owners, clean source data, materiality thresholds, pre-roll simulations, error reports, and formal sign-off. BOMs, routings, scrap assumptions, labor rates, overhead rates, raw material standards, active and inactive parts, sourcing changes, make versus buy, and production line assumptions all need to be reviewed through a controlled process.

You aren’t trying to get to a place where variances are zero. This isn’t an intellectual exercise for finance pros, it’s a tool to drive behavior in the business. You can still budget for variance lines. In fact, you should… standard costing is not your only tool for measuring performance.

One thing I would always watch for is people building “fat” into the standard. Don’t do that. Keep the standards lean and honest, then set a realistic expectation for acceptable misses. It forces a higher level of transparency on production performance issues.

The second bucket is live changes.

These are changes to standard costs that need to happen during the year because the business reality has genuinely changed, or because something new has entered the system. New product launches. Product delistings. New plants. New routings. Maybe a material costing error that is big enough to distort the total margin bridge.

But live changes should be kept to an absolute minimum.

And I mean minimum.

If you are constantly changing standards during the year, you lose the ability to explain what really happened.

Important distinction here: I would correct a material error… but I would be much more cautious about changing a standard just because circumstances changed.

For example, if a product moves from one production line to another and that creates a labor or efficiency variance, I probably want to see that variance. I want to know the real delivered dollar impact of the change. If you immediately encode the new cost into the standard, you may make the variance disappear, but you have also made the operational impact harder to see.

The third bucket is major circumstance changes.

These are rare events where the standard cost base becomes so stale that it stops helping the business understand performance. Examples might include extreme raw material inflation, a major sourcing change, a wholesale production footprint shift, a rapid change in product mix, or a structural change in labor or energy cost.

Your bar for reopening standards mid-cycle should be high.

The test should be: would the performance bridge become materially misleading if we did not change the standard?

If the answer is yes, then fine. You may need a controlled reset. But it should be treated as a proper governance event.

Done well, standard costing updates give manufacturing, procurement, operations, sales, and finance one common language for margin. Done badly, it becomes a giant spreadsheet swamp where everyone argues about whether the problem is real or just “in the standard.”

Thank you for the question. and I truly mean it … thank you.

TLDR: Standard costs should be stable, governed, and explainable. Change them rarely, control them tightly, and learn from variances.

Eli from the US 

We're trying to scale one part of our revenue business 10x over 5 years. Year 1 delivered almost nothing. The core business is contracting unfortunately.

Looking back, I think the problem is we planned like we were still in startup mode—no multi-year roadmap, no clear milestones for when we'd see evidence it's working, no visibility into what needs to happen in Year 2, 3, 4. Accountability and incentives are missing.

Our culture is execution-focused and quarter-to-quarter.

How do you introduce long-range planning function to a scaling initiative when the company has set patterns and cadences that leave a lot of the utility of Long Range Planning on the table?

It sounds like you are at an interesting moment.

In the early days, agility and informality are a tailwind to growth. But as a business becomes more complex, with more priorities, more management layers, and more competing demands, the same things can become a headwind. Suddenly the informality that made you fast, starts making it impossible to get anything done.

I have been in a business that struggled with this transition myself. It is one hell of a cultural reset.

Which brings me to your specific question.

I’m not sure this is a problem you will long-range-plan your way out of.

A 10x revenue goal over five years, by my math, is roughly a 60% CAGR. That is breakout startup-style growth, even if it is trapped inside a bigger business. And if year one delivered almost nothing, I would be careful about assuming the missing piece is a better five-year plan.

If your execution model failed to deliver anything meaningful in year one, making the spreadsheet longer may just make the problem bigger and more beautifully formatted.

This sounds more like an accountability, focus, and resource problem. So I would start by making it smaller, not bigger.

Who owns the 10x goal?

I mean actually owns it. Not “the exec team.” Not “commercial.” Not “growth.” One name.

Then ask what resources that person needs to deliver the next stage. People, product, marketing, technology, capital, sales capacity, management attention. Whatever it is.

And then ask what needs to be true in the next 12 months to prove this thing is working.

“What must be true by the end of year two for us to still believe 10x is possible by year five?” That is a much more useful question for a quarter-to-quarter execution culture, imo.

Maybe the five-year goal is 10x, but the next practical milestone could be 2x or 2.5x in 12 months. That is still hard. But it is close enough for the business to understand, digest, and execute against, while getting you back toward the growth curve.

Then you can break that 12-month target into the real drivers. Leads, conversion, capacity, pricing, product readiness, retention, locations, channels, whatever matters in your business. Put names, resources, milestones and incentives against them.

That is where momentum comes from.

And if your business is comfortable in a quarterly execution rhythm, it should lap that up. Whereas you may struggle to get people excited about year four and year five.

You can still use the longer-range model in the background to understand what the business needs to become, but bring the organization a nearer-term execution frame it can actually use.

The long-range planning point becomes more interesting when it exposes trade-offs. If delivering this 10x initiative requires serious resources, then where is that resource coming from? Is it being diverted from the contracting core business? Is that the right answer?

That is the conversation long-range planning should force.

Over time, as the business gets better at this, you can stretch the horizon. Start with 12 months, then 24, then a proper five-year planning rhythm. But trying to force a full business-wide long-range planning process into a culture that only knows quarterly execution may be like asking a fish to climb a tree.

TLDR: Don’t start with a five-year plan. Start with ownership, resources, and the next proof point.

A few of the biggest stories that CFOs should pay attention to. This is also the section you might not want to see your name in.

The Hugging Face attack is way, way worse than initially thought

I’m not a doomsday prepper by any means, but reading this was enough to send me out to buy a pallet of those giant Costco tins of beans

New CFO takes over at Hershey

This role is going to be a gauntlet of supply chain, cost, and pricing challenges. Nothing is going to fix the dreadful taste of Hershey’s chocolate though (yes I’m a choc snob.)

ICYMI, here are some of my favorite finance/business social media posts from this week.

Let me know what you thought of today’s Mailbag. Just hit reply… I read every message.

My Playbook is back for September, and this month we are covering what to do if you Inherit a Shitshow Finance Function; you can find part 1 here

Disclaimer: I am not your accountant, tax advisor, lawyer, CFO, director, or friend. Well, maybe I’m your friend, but I am not any of those other things. Everything I publish represents my opinions only, not advice. Running the finances for a company is serious business, and you should take the proper advice you need