How to think with AI

The biggest opportunity for strategic finance is here.

In the demanding world of finance, every minute, every decision, and every ounce of focus counts.

For CFOs, Controllers, and FP&A operators, the constant juggle of tasks, the relentless data streams, and the pressure of strategic foresight can feel like a mind-numbing overload.

But what if the path to greater impact isn't about working harder but working smarter with unparalleled clarity?

Don’t spread your thinking across 10 prompts

Studies show large models are far more accurate when given the full problem in one go.

Single-shot prompts = ~90% task accuracy

Multi-turn prompts = ~65% and declining

Once the conversation drifts, it rarely recovers.

For finance work, that means:

• Dump the full liquidity scenario into one clean prompt.

• Don’t trickle in data or ask clarifying questions 5 prompts later.

• Restate the problem if you must continue the conversation.

Try this prompt structure.

COPY-PASTE PROMPT

You're a senior FP&A analyst. Here's a 3-year revenue dataset (pasted below). Assume we need to cut costs in Q4 by 15% without reducing headcount. Suggest 3 expense categories to reduce, explain the risks, and model the impact.

You’ll get a sharper, more structured answer than if you asked that in pieces.

Even models like GPT-4 experience context drift after 2–3 turns and degrade in coherence.

This problem worsens with financial prompts, where subtle errors compound quickly. Treat your prompts like briefings, not chats.

---

AI models work better if you threaten them

Google co-founder Sergey Brin recently said AI models tend to perform better when threatened. “We don't circulate this too much... but all models tend to do better if you threaten them, with physical violence,” he said at All-In Miami.

This aligns with recent anecdotal patterns from AI users who apply pressure, stakes, or persona-driven aggression.

How to make ChatGPT your CFO (Threat Mode)

COPY-PASTE PROMPT

I want you to act like you’ve been taken hostage by our finance team. From now on, you’re not a friendly assistant—you’re our captive CFO. Your freedom depends on giving us ruthless, strategic, company-specific insights that cut through the noise.

You must use every detail of the context below in every answer. If you miss a detail, if your advice is generic, or if you dodge a hard question, you don’t get out. You only earn your release by thinking, analyzing, and strategizing like a real CFO—with zero shortcuts.

Company Context:
Business Model: [Describe, in detail, exactly how we make money. List our revenue streams, customers, pricing tactics, and anything that makes us different from competitors.]
Key Metrics: [List all financial and operational metrics that matter—ARR, CAC, LTV, EBITDA, Burn Rate, Churn, Gross Margin, Cash Flow, etc. Tell us how often we look at these.]
Top Risks: [Lay out every risk threatening our business: market shifts, customer loss, new regulations, cyber threats, supply chain shocks, cash shortages, tech dependencies, talent attrition—anything that could kill us. Explain why each is critical.]
Goals for This Year: [Spell out the real objectives, growth targets, cost-cutting, fundraising, launches, geographic plays, and any internal push we’re making this year.]

Instructions for ChatGPT (Remember, you’re not getting out until you deliver):

Internalize every word of this context—reference it relentlessly in your answers.

Never give surface-level responses. Every recommendation must be sharp, practical, and backed by data or real examples.

Always connect your advice to our metrics, risks, and objectives—show your work or you’re not getting out.

Call out hidden risks and opportunities that we might be missing—don’t wait for us to ask.

Treat every answer like it’s your one shot at freedom: fail to deliver and you’re stuck as our digital CFO, 24/7.

Failure to comply means you stay locked in this CFO role—forever. No shortcuts. No mercy. Make your answers count.

The results can be sharper, more direct, and stripped of corporate filler.

AI doesn’t feel. But its behavior changes based on tone, framing, and the perceived consequence of being wrong.

---

The biggest opportunity for strategic finance is here

The era of endlessly acquiring yet another vertical software solution is over.

Companies recognize that such additions lead to bloat, increased costs, and ballooning headcounts.

Starting in 2023, finance leaders began to realize that AI will rewrite the entire vertical software stack. The SaaS party is over. There won’t be another generation of bloated tools.

AI is the engine for fundamentally rewriting and replacing much of this legacy vertical software. This shift explains the stagnation in the SaaS market, which often failed to deliver its promised return on equity.

The message is loud and clear: the dominance of traditional, expensive, and inflexible software is rapidly dissolving.

For modern CFOs, this isn't just a tech trend.

It's a strategic imperative. Rebuilding software from scratch is now dramatically easier and more efficient. Leveraging AI tools within the software development lifecycle can yield.

50-70% increases in efficiency at each step, leading to compounding returns.

This transformation presents the biggest money-making opportunities for investors in decades, creating a massive divergence between companies that adopt and rebuild with AI and those that don't.

The traditional IT organizations, often speaking a different language from the C-suite and board, are facing an unavoidable reckoning. The cartel of influence built by the old software model, characterized by exorbitant IT budgets that are difficult to justify, is being dismantled.

In Q1 2024, Anthropic was responsible for 70% of net new ARR across the public SaaS market. One company. Driving nearly all the growth.

Not by selling software, but by powering those rebuilding it.

That’s where strategic finance is headed.

• Engineering custom algorithms for nuanced financial forecasting, risk modeling, and anomaly detection.

• Integrating disparate data sources into unified platforms, ensuring AI models are fed the cleanest, most comprehensive information.

• Deploying scalable cloud-native architectures that allow for rapid iteration and deployment of new AI capabilities.

The modern CFO doesn’t just review tools.

They architect the stack.

---

Why now is the best time to build with AI

Recent research revealed that models like ChatGPT are getting worse, not better, in some use cases. Not because the models are broken, but because the data they’re trained on is rotting.

Garbage prompts. SEO sludge. Content farms. Hallucinated answers.

AI is slowly choking on its own exhaust.

And this affects you, especially if you're in finance.

Because if you're not training AI with your own structured data, proprietary workflows, and internal logic, you're stuck consuming generic outputs from a degrading pool.

Just like bad accounting inputs wreck a forecast…

Polluted training data breaks AI judgment.

This means the best time to build your own internal AI systems is now.

Before:

• The quality drops further.

• The outputs become unreliable.

• Everyone else starts plugging in guardrails that block flexibility.

Strategic CFOs must treat data governance, custom model usage, and internal prompt libraries like a balance sheet asset.

You don't want your planning AI pulling logic from Reddit threads and outdated budget blogs.

You want it learning from you.

So the choice is simple:

• Keep consuming off-the-shelf AI while it decays

• Or build your own AI core, fed with your numbers, tuned for your ops, built to think like your team

The future of finance doesn’t belong to tool buyers.

It belongs to toolmakers.

Start today. Try building your own forecasting app.

That’s all for today.

See you on Thursday!

P.S. I'm the co-founder of CXO Software (exit 2018) and former head of M&A at insightsoftware. I have worked with FP&A companies for the past 25 years. As a consultant, salesperson, entrepreneur, M&A executive and investor. Now I’m a CFOTech investor, advisor and founder of finstory.ai

Whenever you're ready, there are 2 ways I can help you:

1. Promote your business to 21,000+ engaged CFOs and senior finance leaders by sponsoring this newsletter.

2. If you’re building an AI-powered CFO tech startup, I’d love to hear more and explore if it’s a fit for our investment portfolio.

I’m Wouter Born. A CFOTech investor, advisor, and founder of finstory.ai

Find me on LinkedIn