Accounting is a solved problem. Debits equal credits. The rules are defined by standards bodies. Yes, AI will likely take over accounting in the next several years. The foundations are there to translate documents into debits and credits, make payments, post journal entries, and complete daily reconciliations. An accountant will press a button to confirm (at least for a while). If you don’t believe me, download the IFRS or GAAP standards, store them in ChatGPT, use Mistral to OCR invoices, and have ChatGPT give the exact journal entry with dates. Give it something complicated.

Planning, on the other hand, is trickier. What should your strategy and budget be for next year? What should your company do tomorrow? These aren’t solved problems. These require decisions based on context and unquantifiable information. And although gen AI can currently provide some guidance, I would put it up there at the same level of helpfulness as downloading OpenCapital’s SaaS benchmarks and saying, “Now, if we just get CAC down to $7,000 and up prices to $14,000, we will have a best-in-class CAC payback to raise our next round. We should cut marketing costs and raise prices.” Mic drop.

For better or for worse, forecasting and budgeting will stay in our desperately human FP&A hands. Tomorrow has too many opportunities. Rallying a team in one direction involves trade-offs, taking informed risks, and dealing with unknowns. This applies to nearly every decision from how you price to when you are going to hire to whether your investment will pay off. Running through scenarios is easy (hopefully at this point you have an automated solution for this), but what matters is the path your company chooses to follow.

Twenty years ago, in a school valuation challenge, we built detailed operational models, identified key variables, and ran Monte Carlo simulations. That gave us nearly every possible financial outcome for their next five years. Guess how helpful it was… The winning team ran the numbers, but spent most of their time analyzing the market, talking to people, and peering out into the real world. That will always be the winning team.

Don’t misunderstand, AI is great. If you're not using it now, you are already behind. Use it to identify data errors. Use it to brainstorm, produce policy guidelines, train your team, fix typos, and insert AI into every fifth word of your investor pitch. Get the rush of sinking hours into writing massive Python files that save five minutes of your time.

But the hard part will always be making the right decision. Justifying the path you have chosen. Using numbers to navigate an uncertain future. Holding yourself accountable. That’s FP&A’s job. Not to present the “path of all possible options” but to gather all the information into one place and steer the ship through the unknown.

And yes. That takes planning. An agreement on who is going to do what in pursuit of what goal. (whisper) A budget.

So, as an FP&A leader, embrace AI. Build your systems around it. But it won’t make decisions for you. No, fortunately, that is your job.