Kairos Power

Eyal Greenberg, CPA (Isr.), MBA, M.Sc., is Director, Finance at Kairos Power, an advanced nuclear technology company, where he leads enterprise planning, forecasting, performance management, executive reporting and finance process improvement. He has more than 20 years of experience across strategic finance, FP&A, financial modeling, finance transformation, analytics, M&A advisory and automation, including leadership roles at PwC, RSM, KPMG and EY. His work focuses on building scalable finance processes, improving decision quality and translating financial and operational data into actionable insights for executives, boards and cross-functional business leaders.

In many finance transformation discussions, the initial request sounds simple: automate the forecast. But when the process is reviewed more closely, the real issue often becomes clear. The company does not just need automation. It first needs a standardized forecasting process.

That distinction matters. FP&A automation is often treated as a technology project. New planning tools, dashboards, workflows and AI capabilities can make it feel like the tool is the solution. It is not. Automation only works when the process behind it is clear, repeatable and disciplined.

Automating an unclear process does not create efficiency. It creates a faster version of the same confusion.

Finance teams should not start by asking, “What can we automate?” They should start by asking, “Which parts of our forecast are consistent enough to automate?” If the answer is unclear, the organization may not yet have an automation problem. It may have a process definition problem.

There is real value in automation. It can reduce repetitive work such as data collection, transformation, mapping, reconciliations and reporting. These tasks take a lot of time during the FP&A cycle, but they rarely improve the quality of the final business decision. When automation removes this manual work, finance teams can spend more time on analysis, business partnering and decision support.

But automation can also become a burden. If the forecasting logic changes every cycle, the automation will need constant updates. If every department submits inputs differently, the automation will only expose that inconsistency. If key assumptions live in someone’s head or in offline spreadsheet tabs, the process is not ready.

The first deliverable of an FP&A automation project should not be a dashboard. It should be a clearly documented forecasting process.

That is one of the most useful benefits of automation work: it forces the organization to be specific. Who owns each input? What is the source of truth? What logic is used? Which assumptions are recurring? Which items require review? Many forecasting processes work for a while because experienced people know how to make them work. But as the company grows, informal knowledge becomes a risk. Automation forces that knowledge into the open.

This is why FP&A automation is not only a technology effort. It is a finance transformation effort.

A practical way to approach it is through a segmented approach. Not every forecast item should receive the same level of automation. Some items are recurring and predictable. Office rent, recurring software costs, routine operating expenses and other stable cost categories can often be forecast using consistent logic, contract terms or historical patterns.

Automating an unclear process does not create efficiency. It creates a faster version of the same confusion.

Other items should not be placed on autopilot. Large onetime expenditures, government grants, legal claims, licensing costs, unusual vendor activity and other non-recurring items often require context and judgment. Automation can still help track these items, organize the workflow, document assumptions and improve visibility. But the forecast itself should not blindly follow a standard formula.

A simple test is if a line item cannot be explained consistently by the finance team, it is probably not ready to be automated. The danger is false precision. Once a forecast is automated, the output can look more reliable than it really is. It appears structured. It appears consistent. It may even appear more sophisticated. But finance leaders still need to ask basic questions. Does this reflect the current business reality? Are there one-time expenses that should be separated from the run rate? Has something changed that historical logic will not capture? Are we reviewing the assumptions or simply repeating them because the system produced them?

AI can help FP&A teams identify trends, summarize variances and flag anomalies. But AI does not remove the need for finance judgment. In many cases, it makes judgment more important. Finance teams still need to decide which outputs matter, which assumptions need review and which exceptions require business context.

The goal should not be fully autonomous forecasting. The goal should be a more disciplined, transparent and scalable finance process. Automation should handle repeatable work. Finance professionals should focus on assumptions, exceptions, tradeoffs and decisions.

The organizations that get the most value from FP&A automation will not be the ones that automate the most activity. They will be the ones that know what should be automated, what should be reviewed and what still requires human judgment.

FP&A automation is not just about faster reporting. It is about building a finance function that supports better decisions with more consistency, transparency and discipline.