Why CFOs Can't Delegate Their Voice to an Algorithm
Pedro San Martín — Principal at Asher PwC Interamericas
The quarterly board meeting was going smoothly. Clean slides, solid margins, targets met. Then an independent director cut through the comfortable silence: "What actually changed in customer behavior to explain this sudden improvement?"
The room went quiet. Our CFO's report was full of the usual suspects—"operational efficiency," "process optimization"—but it didn't answer the real question. It couldn't, because it had been written by a machine.
The CEO leaned over and whispered what we were all thinking: "This was auto-generated, wasn't it?"
That moment crystallized something I've been seeing across Latin America's finance departments: we're losing our voices to our own tools.
When Reports Stop Being Reports
Here's the thing about modern finance leadership—the challenge isn't getting the numbers anymore. Any decent EPM system can spit out variance reports faster than you can say "EBITDA." The real challenge is making those numbers mean something.
In our 2024 survey of 128 CFOs across the region, three-quarters reported that their stakeholders won't make decisions based solely on dashboards. They need the story. They need context. They need someone who understands the business to connect the dots.
However, somewhere along the way, we began treating narrative reports as an administrative task rather than what they truly are: the written embodiment of our professional judgment.
A retail CFO in Panama put it perfectly: "The board doesn't just want the number—they want to know I understand what it means."
The Seductive Efficiency Trap
I get why AI is tempting for financial reporting. It's fast, consistent, and never misses a deadline. But here's what Harvard's 2024 research showed us: people inherently trust human-written content more, even when they can't tell the difference.
We've tested this ourselves. In workshops across Florida and Central America, we presented finance leaders with two explanations for the same budget variance. Identical accuracy, similar length. But 67% trusted the one that "felt more human"—even though both were technically correct.
The real danger isn't that AI writes poorly. It's that it writes adequately. And adequate financial communication is like adequate surgery—it might not kill you, but it won't heal you either.
I've seen reports that explain margin improvements as "SG&A efficiencies" without providing context, references to actual decisions made, or acknowledgment of trade-offs. Technically accurate. Strategically useless.
Three Rules for Reclaiming Your Voice
After working with dozens of CFOs who've wrestled with this balance, here's what actually works:
- Start with AI, finish with you. Let the machine draft the structure, catch the obvious patterns, flag the outliers. But the final interpretation—the part that says "here's what this means for our business"—that has to come from you. A sentence like "these margin gains reflect some difficult choices we made last quarter" carries weight that no algorithm can generate.
- Know who you're talking to. Your investor memo can be technical. Your board update needs strategic context. Your team communication deserves empathy. Tone isn't just style—it's leadership. And leadership can't be automated.
- Show your thinking, not just your conclusions. Strong financial narratives don't just explain what happened—they explore why it matters and what comes next. Instead of "Business Unit X grew 12%," try "Business Unit X grew 12%, but client churn jumped 18%—we need to understand this disconnect before celebrating."
Technology as Partner, Not Replacement
This isn't about rejecting AI. Smart CFOs are using it to draft text, spot inconsistencies, and suggest structures. But they're not letting it have the final word. As one telecom CFO told me: "I use AI as a starting point, never as a finish line. The voice of Finance should sound like us."
The best financial leaders I work with treat AI like a very efficient research assistant—helpful for the heavy lifting, but not trusted with the insights that matter most.
The Test That Matters
Before you send your next financial report, ask yourself: Could this have been written by any other CFO at any other company? If the answer is yes, you're not done yet.
Your stakeholders have access to the same data you do. What they're paying you for is judgment, context, and the courage to say what needs to be said—even when it's uncomfortable.
In an age of infinite data and auto-generated dashboards, the real differentiator for CFOs isn't showing numbers—it's giving them meaning. And that still requires something no algorithm can provide: understanding your business well enough to explain why the numbers matter.
AI can suggest words. But strategic narrative remains human. Every report should make the reader to think: "This message comes from someone who truly understands our business, not just our systems."
Pedro San Martín
Principal at Asher PwC Interamericas
psanmartin@asheranalytics.com
References
- Choudhury, P. et al. (2024). The Wade Test: Generative AI and CEO Communication. Harvard Business School Working Paper
- Deloitte (2024). CFO Signals Survey – AI Adoption in Finance
- PwC (2024). Finance Effectiveness Benchmarking Study
- PwC (2024). Pulse Survey: Finance AI Adoption
- PwC (2024). Responsible AI Survey
- Wharton (2024). The CFO as Strategic Partner
- Wharton (2024). AI in Finance: The Promise and Pitfalls