Can AI help with capital allocation?

Ask Claude where AI tools can help with capital allocation, and Anthropic’s LLM says it can read 10-Ks and other research, build and stress-test discounted cash flow models, lay out options, play devil’s advocate, and calculate the per-calorie cost efficiency of donuts vs. bagels. Ahem, sorry, that last one was us trying to stretch our team meeting eats budget. But you get the picture.

“Think of it as a fast, well-read analyst who can do a first pass, challenge your thinking, and catch mistakes,” Claude said in its response.

That sounds good, but Claude also warned that it “shouldn’t have final say.”

CFOs are being asked to incorporate AI into as many areas of the business as possible, but some functions are more AI-ready than others. Capital allocation decision-making could be one of those finance settings where CFOs aren’t yet ready to trust AI.

According to Esker’s 2026 Global Finance AI Trust Index, fielded in August, 24% of 3387 finance leaders across the US, Canada, UK, EU, and Australia said they’re prioritizing AI investment in capital allocation in the next year. But that was behind planning and forecasting (51%), financial reporting (42%), liquidity and cash flow management (39%), and invoice processing (30%), according to the business process automation company.

Esker CFO Scott McDermott told CFO Brew via email that the survey results make it clear that AI is becoming more trusted, but that it “does not necessarily extend to fully autonomous decision-making yet.”

“[Trust] increasingly extends to recommendations, analysis, and providing context to support those decisions,” he said.

Echoing McDermott, IBM’s 2026 CFO Study also reported that “AI is increasingly trusted to execute lower-risk activities, but confidence drops sharply as decisions become tied to capital allocation, fiduciary responsibility, and strategic exposure.”

Metrics vs. nuance. Michiel Boere, CFO of HR software platform Remote, stressed the need for human judgment in tandem with any AI-enabled decision.

“Can AI help allocate capital? Well, if you ask it, it’ll do a very confident capital allocation. And it’ll argue with you that it’s right,” Boere mused.

“There is judgment that comes into where you should spend your money, and AI can give you advice like any consultant will give you advice. It sounds very convincing, but you have to make up your own mind on whether you actually believe it.”

In addition, “some capital allocation decisions are relatively straightforward and highly analytical…Other capital allocation decisions are far more nuanced,” McDermott said. “Evaluating an investment using metrics such as payback period, NPV, or IRR is largely a data and modeling exercise, and AI is well positioned to help accelerate and improve those analyses,” he said.

“[Nuanced decisions around allocating capital] may involve choices around dividends versus buybacks, acquisition opportunities, hiring plans, product investments, marketing spend, or determining where the next marginal dollar should be deployed across the business,” McDermott said. “Those decisions often require context, judgment, and a deep understanding of company strategy, competitive dynamics, and management priorities that do not fully reside in enterprise systems.”