Fora scaled 3x in a year. Their finance team didn't.
Join Campfire, Ramp, and Fora's Controller for a behind-the-scenes look at how Fora managed millions in spend and hundreds of thousands of transaction lines per month while cutting close time by 20%.
You'll walk away with the exact stack they built, where they automated, and what it really takes to run lean at hypergrowth speed.
🚨 ANNOUNCEMENT 🚨
Are you considering a new ERP implementation? Or maybe want to understand better what an AI-native ERP actually means?
I want to create more content that can help you with that. Take this two minute survey to tell me your biggest questions about implementing a new ERP solution.
Now, on to today’s Mailbag…
Landed in a tight corner? Shoot me a line (anonymously if you wish), and you might see me answer your question below.
👉 Send me your questions by filling out this form.
We’ve got some great topics. Here’s what’s on tap:
Integrating AI into existing financial models
Managing an owner’s monthly distributions
Striking a balance between wartime and peacetime
Now, let’s get into it.
John from the US asked:
I am in charge of my company’s financial model (aka "The Model"). It is a model I built from scratch that we use for decision-making (e.g., stress testing, scenario planning, M&A capacity, covenant calculations, etc.)
I’ve played around using Claude to run DCFs, summarize 10-Ks, and perform other tasks, but I don’t see how it can help with The Model itself.
How do I get started implementing AI?
Thank you for the advice!
John, you've hit on something important here.
Most AI content focuses on building something new from scratch. A model, a dashboard, a tool. But most finance teams will have things that already work.
The more interesting question is how you fold AI into existing workflows without breaking what you've built. Or at least so you can choose what you break.
Let me start with what I would not do.
Do not rebuild the model for the sake of it. If it works, it is a mission-critical decision-making tool. Your job is to make sure it is there and working when the business needs it, not to run needless experiments on it.
And do not Claude Code it into a shiny app with sliders and visualizations just because you can. Excel is the perfect thinking canvas for finance. Rows and columns are the right way to think about the kind of use case you describe. An app wrapper with fancy sliders would not make it better. It would make it less transparent.
If you were building this from scratch today, then I would say use Claude for Excel. I have experimented with it a fair amount, and it is genuinely good at building financial models, provided you know exactly what you want and can brief it the way you would brief a sharp analyst.
Where it struggles is iterative development, giving it incremental instructions on top of itself over time. You end up with ugly, nested, overcomplicated formulas that technically work but are nearly impossible to audit. Not what you want in a mission-critical model.
But what would I actually do with an existing model?
Integrate slowly, around the periphery, and build confidence before you let it anywhere near the core.
Here are some things I’ve played with to some success:
Audit the model. Ask Claude to review your existing model and produce a separate tab detailing potential vulnerabilities, broken references, circular logic, or assumption inconsistencies. While you are at it, have it document the model's logic and assumptions in plain English. Genuinely useful for onboarding new people and a good test of whether the model actually does what you think it does.
Help with the actuals rollover. Every month or quarter, you are rolling forward with new actuals. Can Claude help sense-check whether your forward assumptions still hold, given the new data? Start with it as a critical friend before you let it work as an agent.
Run a wider scenario set. This is where Claude's computation capability really earns its keep. Run a Monte Carlo simulation, letting it vary inputs across a high number of scenarios without touching the formulas, and produce a probability distribution of outcomes. That kind of analysis is hard to do manually and genuinely useful for stress testing and covenant headroom conversations.
Build a rebuild prompt. Keep a separate tab where Claude maintains a running "one-shot" prompt that would theoretically reproduce the model as built. Then open a fresh workbook and test that prompt. Did it work? This is both a great audit tool and a forcing function for model documentation.
One practical note: if Claude starts touching even peripheral parts of the model, make sure you have a clear version control discipline in place before you start. Mission-critical tools need a clean change history.
The theme across all of these is using Claude's computation and analytical capability to do things that are hard to do manually, without breaking the things you have confidence in. You will soon start to see the natural opportunities to let it do more.
TLDR: Don't rebuild what works. Integrate AI slowly around the edges, use it as a critical friend first, and let it earn the right to touch the core model over time.
Austin from the US asked:
Hi, I’d love your perspective on a situation I’m facing as Head of Finance.
I work in a privately owned company where the owner takes regular distributions regardless of actual cash generation, even in flat or loss-making months. While the business may look fine from a P&L perspective, this creates ongoing pressure on cash and gradually erodes our liquidity buffer.
In strong periods, the business still generates enough cash to offset distributions and build reserves. But when margins tighten, or the business is flat or losing money, distributions continue unchanged; this creates cash constraints, limits flexibility, and leaves us exposed to unexpected events or missed growth opportunities.
The challenge isn’t a lack of understanding— the difference between profit and cash flow is clear. The issue is that distributions continue anyway until the situation becomes visibly critical.
How would you approach this? Have you seen any practical guardrails or mechanisms that actually work to align distributions with real cash generation, especially when you don’t have final decision authority?
Austin, it sounds like your owner may be a disciple of Profit First by Mike Michalowicz.
That book (focused at owners) flips the traditional equation. Instead of Profit = Sales minus Expenses, it argues that Expenses should equal Sales minus Profit. It acts as a forcing function on the business, teaching it to ‘eat from a smaller plate.’
It is a slightly alien idea for finance people, but honestly, I don't hate it for owner driven businesses…
I have worked with an owner who implemented something similar, rigidly and without exception. It absolutely changed behavior in the business, for better and for worse. Businesses exist to deliver a return, and a fixed distribution is a blunt force instrument to making sure that either happens, or is flagged very quickly.
It is also worth understanding why owners do this, because it is rarely as irrational as it looks. Often it comes from entrepreneurs who have been burned before or who simply do not trust accruals-based accounting. The owner I knew didn't fully trust what the P&L was telling him, so taking cash out each month was his way of stress-testing the gap between reported profit and actual cash generation. Not how we are taught to think about it, but not everyone wants to run their business from the financial statements.
There is a reality check here… this is a battle you are probably not going to win. An owner who has been taking a fixed monthly distribution for years will sooner give up something far more precious than that habit.
So the short answer is: don't fight it. Work with it. Then try to influence it over time.
Build the distribution into the budget like any other fixed cost. Show the owner you are treating their return as seriously as they are. Then, when a difficult month does arrive, an owner who believes you are on their side is far more likely to listen to you.
Then step back and ask whether the level is actually sustainable over the long term. You mention periods where reserves build and others where they deplete. Is the long term average annual take a reasonable return on the business relative to your competitive set and industry? Or is this more of a cash flow phasing problem than a structural one?
If it is phasing, there may be room to negotiate some shape into the distributions without changing the annual total.
Smoothing them to reflect seasonal working capital cycles, for example, or adjusting the timing around known capex or tax payments. Small changes like that can meaningfully reduce the pressure without asking the owner to give anything up in aggregate.
If you get some engagement, then set it all out very clearly in a schedule so they can see what they will get and when, and show them you can deliver exactly what they expect, but in a way that is healthier for the business.
If you are seen to be leaning toward the owner rather than away from them, you have a much better chance of getting them to move. Gradually and on their terms, but move nonetheless.
If growth is genuinely being stifled, put that to the owner directly and help them see the trade-off clearly. It is ultimately their choice, but they should make it with full information.
One caveat to all of this is that if the business carries any debt, be aware that persistent distributions during loss-making periods can start to raise creditor preference questions. Worth knowing where that line is.
TLDR: Don't fight the distribution. Plan to deliver it, understand whether the level is sustainable, and negotiate shape and timing rather than the total. Move toward the owner, and they are more likely to meet you halfway.
John from the US asked:
I thrive in wartime, but my business partner thrives in peacetime. How can we co-exist over the long run?
For background, I am a VP who specializes in distressed situations. We hired a second, peacetime VP once the department stabilized, and I struggle with all the recognition they receive for advancing the team forward.
Can a wartime and peacetime employee work side-by-side?
I like the wartime and peacetime manager concept. It is a useful lens for thinking about turnarounds and what different phases of a business actually need.
But I want to push back on it a little before I answer your question directly.
The framework can be misused. Peacetime manager is sometimes just a polite way of saying someone is a do-nothing who is out of their depth. And wartime manager can be a flattering label for someone who is simply unnecessarily disruptive.
I also think it is a little ‘2010s’. A period of relatively modest macro peaks and troughs. This decade has been something else entirely. One black swan event after another until we have a full flock. Frankly, every leader needs a little wartime.
And I genuinely believe the old cliché: hard times create strong managers, strong managers create good times, good times create weak managers, and weak managers create hard times. Round and round.
What we are watching play out in the SaaS market right now is a generation of industry leadership that had it so good for so long, they never built the muscles to pivot or to see the damage coming before it was too late.
Anyway… back to your situation.
Can wartime and peacetime employees coexist? Of course they can. In fact, the best teams will likely have a bit of both. My style is that I like wartime more than peacetime, but I like situations where there is a real job to be done. I don’t want to babysit someone else’s work.
I think you need to honestly consider two possibilities here.
The first is that your peacetime counterpart is genuinely skilled at something you undervalue.
Making things look easy is a real skill. Stabilizing a team, building process, retaining people, and creating a culture where good work compounds quietly over time. That is not nothing. It might not feel as viscerally satisfying as a turnaround, but it creates real value.
The second is simpler. Maybe they are getting credit for work you did. If that is the case, you probably did not play the game quite right.
Pure wartime leaders (and if you are a special sits person this is probably you) often make the mistake of moving on to the next fire before the story of the last one has been properly told.
Twelve months leading a department through stability, after you rescued it, would have made that credit much harder to reassign.
Either way, the answer points in the same direction. Learn to run the full cycle. Not just the disruption, but the stabilization that follows. If you want to be a great CFO one day, that is the job. Taking complex, difficult things and digesting them into simple things the business can run easily.
If you can only thrive in the chaos, you are half the leader you could be. And at some point, that ceiling will show up.
The recognition problem solves itself when you own the whole arc, not just the exciting part of it.
TLDR: Wartime skills are valuable, but guaranteeing peace is harder than you think. Learn to run the full cycle, and the recognition problem takes care of itself.
A few of the biggest stories that every CFO is paying close attention to. This is the section you might not want to see your name in.
Being a hardware CFO is harder than being a software CFO. There are simply many more moving parts. Interesting to see Oracle seeing that gap and recognizing it head on. Will we see more tech companies employing more robust CFOs who understand inventory and capex? I think so.
Physical inventory count is the grunt work that built generations of auditors. And it's still doing it. In fact, I’d bet that physical inventory audits is absolutely the last thing AI will take out of accounting.
A CFO at a pipeline company in Texas spent years using company money on luxury goods and vacations before anyone noticed.
ICYMI, here are some of my favorite finance/business social media posts from this week.
I created SimCFO, a hands-on finance simulation tool where you get to practice thinking like a world-class CFO. You spend 12 months in the Interim CFO seat of a struggling sports apparel business to steer it out of the smelly stuff.
And at the end, you’ll get to see if you bagged the job permanently (and even get to negotiate your comp.)
I’m not saying it’s a contest, but also … yes, it’s a contest. So here’s this week’s leaderboard of folks who completed SimCFO. Congratulations to George Sharp who is currently top of the pops…
If you’re looking to sponsor CFO Secrets Newsletter, fill out this form, and we’ll be in touch.
If you enjoyed today’s content, don’t forget to subscribe.
You can help make sure this newsletter always stays free simply by spreading the word. And when you share CFO Secrets with your finance friends, you’ll earn rewards, including a 50-page PDF guide on what it takes to be a great CFO. Start sharing your unique referral code today: {{rp_refer_url}}
Let me know what you thought of today’s Mailbag. Just hit reply… I read every message.
Don’t forget to take the two minute survey to tell me what new content you’d like to see on ERP software.
Last weekend’s Playbook explored how automation and AI are different and when to use each.
In last week’s Boardroom Brief, how has the war is affecting 2026 refinancing decisions.
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
