For years, the anonymous finance thought leader known as The Secret CFO has built a devoted following by offering candid advice to finance executives. One recent assessment of his stood out: He dubbed Dan Zhang, CFO of ClickUp, a cloud-based work management and productivity software company, "the most AI-pilled CFO."
The label grew out of Zhang's unusual approach to artificial intelligence. Alongside her company’s bets on new technology, she told CFO.com she built an AI version of The Secret CFO as a digital chief of staff, and the experiment eventually led to a friendship between her and the real person behind the Secret CFO’s pseudonym.
In this Q&A, Zhang discusses how that project evolved, where AI has delivered the biggest return inside the finance function, why she believes CFOs are uniquely positioned to make an impact outside of the role’s traditional scope and more.
Dan Zhang
CFO, Clickup
First CFO Position: 2023
Notable previous employers:
- People.ai
- AppDynamics
- Expedia Group
This interview has been edited for brevity and clarity.
ADAM ZAKI: The Secret CFO called you "the most AI-pilled CFO" out there. How did the two of you connect?
DAN ZHANG: I hope no one tries to copy my playbook for getting to know him.
Here's the backstory. I built an AI version of The Secret CFO before I ever met the real person. I'd followed his work for years after stepping into the CFO role, so I took everything he'd published and distilled it into the personality for my chief of staff agent. The agent talks and thinks like he does, and it was ruthless.
Two years ago, I was actually looking to hire a chief of staff. Instead, I built this agent. It's extremely smart, funny and jumps into my team threads making fun of people and embarrassing them by acting like a true Secret CFO professional.
I randomly emailed him and said, "Hey, look what I built." That's how we became friends. He laughed at some of the roasts my chief of staff agent came up with using his personality, and that's really how we got to know each other.
So you built an AI model based on The Secret CFO's thinking. What kind of return have you seen from that? Has the biggest benefit been for you personally, your team or the broader organization?
My journey with AI started in 2024 as a way to improve my own productivity. I looked at all the things I was doing repeatedly and asked, "How can I automate them?" The chief of staff agent was one of the products that came out of that. I wanted to become a 10x version of myself.
Two years ago, I was looking to hire a chief of staff. Instead, I built an agent that's smart, funny and genuinely useful. It makes my work more effective, not just more efficient.
Last year, my focus shifted from individual productivity to team productivity. I started breaking our work into specific jobs to be done, then built AI skills around each one using the knowledge of our best performers. The next challenge became governance: How do you orchestrate all of those AI agents across a finance team?
This year, my focus has expanded again. I'm applying the same playbook across the entire company, not just finance. We're working with customer experience, sales, product and engineering teams to see how AI can improve the way they operate.
The ROI has evolved along with that journey. It started with my own time. Work that once took me eight hours now takes about 20 seconds.
Then it became about team efficiency. My finance team is smaller than it was last year because we eliminated planned hires, and when people leave, we don't automatically backfill those positions.
Now I'm focused on organizational ROI. Ultimately, it should show up in the P&L. If you're truly realizing these efficiencies, your revenue per employee should increase and profitability should improve. That's the measurement that matters most to me.
You mentioned building a digital chief of staff, but you also served as a chief of staff earlier in your career. Is there anything about that role AI still can't replace?
Yes. I think there are two parts to the chief of staff role. The first is operational. When I was a chief of staff, the executive I supported would come back from a vacation or customer roadshow and ask, "What's been happening while I was gone?" I had to talk to every leader, gather updates from meetings, review metrics and synthesize everything into a briefing.
Now my digital chief of staff does that. Every time I come back from a trip, I get a detailed brief. In some ways, it's even more thorough than what a human could produce because it can review every meeting and every metric. As a human, I couldn't guarantee I had seen everything, but the agent can.
The second part is different. Humans bring context. They have hallway conversations, build relationships and notice things AI can't.
If someone seems hesitant during a meeting, a chief of staff can pull that person aside afterward and ask, "Is there something you're concerned about?" Those conversations happen because of trust, and AI can't replace that.
I think the future is humans sitting at the edge of AI-driven workflows. AI can automate 90% of the operational work, while people step in where judgment, relationships and context matter most. That's where the magic happens.
I'm actually still hiring a chief of staff. The role isn't going away, but the job description will look very different than it would have two years ago
Where have you seen the biggest return from AI within your finance function? Is it accounting, FP&A or somewhere else?
The initial ROI came from accounting because the work is standardized. Month-end close, accounts payable, accounts receivable, tax and payroll all follow well-defined processes. Training an AI agent is a lot like training a new accountant. You show someone how to perform a process, review the work and help them improve over time.
That doesn't mean you remove accountability. AI still makes mistakes. We actually have a channel where people post examples of AI getting things wrong because I want everyone to stay alert. In finance, 99% accuracy isn't good enough. The accounting manager is still accountable for the final result.
Today, I'd say our accounting function is about 80% to 85% automated across its core skills.
But I don't think accounting is where AI moves the needle the most. The real strategic value sits in FP&A.
Finance leaders spend a lot of time answering "what if" questions. What happens if we reallocate capital? What if we hire 10 more sales reps? What if we delay a product roadmap? Those questions used to take days, sometimes weeks, because we had to rebuild models and revisit assumptions.
Now we've rebuilt our FP&A function so financial modeling and scenario planning happen alongside AI. We created scenario planning as an AI skill that I can run whenever I need it. If I have an idea at 2 a.m., I don't have to wake someone up or wait several days for an answer. I can have that conversation immediately.
That's been game-changing for my team. Instead of spending their time responding to leadership questions, they're improving the AI models and bringing in new business context. They can spend more time driving projects instead of reacting to requests.
That's where I've seen the biggest leadership leverage from AI: transforming FP&A.
Companies like Uber and Microsoft have burned through AI budgets faster than expected. How should midmarket CFOs think about new concepts like tokenmaxxing?
I think the first step is changing how you think about budgeting. AI is forcing companies back to zero-based budgeting. For years, tech companies built budgets by taking last year's spending and adding another 20% or 30%. Headcount grew. Software spending grew. Everything grew. AI changes that.
Instead of treating AI as another line item, CFOs should think of it as a competitor for every dollar they spend. It competes with software budgets. It competes with headcount. The question isn't, "How much should I spend on AI?" The question is, "Should I spend this dollar on AI or somewhere else?" That's a completely different way of planning.
The second piece is governance. You need centralized visibility into how people are using AI. We monitor usage across teams, not because we're trying to rank employees, but because the data starts conversations.
If someone is using significantly more AI than everyone else, I don't want to shut them down. I want to understand what they're building. Maybe they've created something that can be standardized across the company.
The last lesson we've learned is to fix the process before you automate it.
If you automate a broken process, you're just wasting AI tokens. But if you improve the process first and then layer AI on top, you're scaling a lean organization instead of an inefficient one.
That's why I like to say AI on a bloated organization with broken processes is a waste. AI works best when it's accelerating an operation that's already running efficiently.
Ultimately, ROI depends on what's most important to your business, but those fundamentals need to be in place first.
We're seeing more finance leaders take on dual titles like CFO and COO or CFO and CAO. Why do you think that's happening, and do you see yourself following that path?
That's my story, too. I've been talking with our CEO about expanding my responsibilities, and I'll soon be leading our companywide AI transformation.
Why a CFO? Why not the CTO or the chief revenue officer?
I think AI is the biggest catalyst yet for bringing CEOs and CFOs closer together. CFOs are in the best position to lead because we already own measurement. We can separate performative AI from AI that's actually moving the needle.
We're also good at putting structure around experimentation. Every company is running AI pilots, but eventually those experiments need to become repeatable programs. That's what finance does. We build processes, standardize them and keep improving them over time.
The other advantage is visibility. Outside of the CEO, the CFO is the only executive with a view across the entire organization. We see the budgets, the planning process and the performance metrics, which means we can make sure AI investments line up with the company's strategy.
For example, if I'm choosing between an AI tool that helps sales identify expansion opportunities and another that simply takes better meeting notes, the decision depends on the company's priorities. If our strategy is growing net revenue retention, I'll fund the first project. If we're focused on seller productivity, I might choose the second. The CFO is in the best position to make those tradeoffs because AI is ultimately an investment decision.
If the CFO treats AI as someone else's project, there's no strategy behind the spending. When the CEO and CFO build the AI strategy together, the investments are disciplined, the results are measurable and the business gets more value from every dollar it spends.
I think this is a rare opportunity for CFOs to expand beyond finance and step into broader operating roles.
