■ CRITICAL RISK ■ Finance
The spreadsheet-and-reconciliation core of the job, yes — treasury management systems and AI forecasting are absorbing it now. What remains is a smaller profession focused on strategy, banking relationships, and the crises models don't see coming.
“Cash flow modeling AI doesn't need Excel or anxiety.”
Our AI replacement risk score — how we score jobs
Treasury analysts manage a company's cash: building daily cash positions from bank feeds, forecasting inflows and outflows, moving money between accounts, initiating investments of surplus cash, monitoring debt covenants and FX exposure, and reconciling everything. The workday is famously spreadsheet-bound — downloading balances from a dozen bank portals, stitching them into a position sheet by 10am, then updating the 13-week cash forecast that leadership will question anyway.
Treasury management systems already automate the plumbing: API bank connectivity kills the portal-hopping, cash positioning assembles itself, and payments route with rules-based approvals. The AI layer targets the analyst's crown jewel — forecasting. Machine-learning models trained on historical flows, AR/AP data, and seasonality produce cash forecasts that beat hand-built spreadsheets, and they update continuously instead of weekly. Anomaly detection flags fraud and errors in payment runs. Even FX hedging execution is increasingly rule-driven. The 'analyst' part of the title — collecting, assembling, projecting — is precisely the automated part.
What survives is treasury as judgment: capital structure decisions, negotiating credit facilities and bank fees, designing hedging strategy rather than executing it, managing liquidity through a crisis when models trained on normal times go blind, and the internal advisory role of explaining cash reality to CFOs. Those live at senior levels — treasury manager, assistant treasurer — and the automation squeeze specifically thins the junior analyst seats where careers used to start. Our 80 risk score reflects a genuinely useful function whose entry-level workload is being deleted, leaving fewer, more strategic chairs. The function thrives even as the traditional entry job quietly disappears beneath it.
Automatability: our editorial assessment of current and near-term AI capability
Adoption is happening now: API bank connectivity and AI cash forecasting are standard features of modern treasury platforms, and mid-size companies are implementing them this decade, not next. Junior treasury analyst openings will keep thinning through the late 2020s as the daily-positioning workload disappears. Senior treasury roles hold steady — someone still owns liquidity risk — but the pyramid loses its base.
It's a narrowing doorway. The entry-level tasks that trained past generations — daily positioning, forecast maintenance — are exactly what treasury systems now automate, so junior seats are shrinking. If you enter, join a team implementing modern tools, learn the systems deeply, and sprint toward risk and strategy work rather than settling into reporting.
For routine operational forecasting, generally yes — models digest AR/AP history, seasonality, and payment behavior at a granularity spreadsheets can't, and they update continuously. Where they fail is regime change: a crisis, an acquisition, a customer collapse. Human judgment about what the model can't know remains the analyst's residual edge.
Bank relationship management and fee negotiation, hedging and capital-structure strategy, liquidity crisis management, and the technical skill of running and auditing the automated treasury stack itself. Certifications like the CTP still signal seriousness, but pairing them with data skills is what distinguishes the survivors.
Yes, pragmatically. SQL and Python won't make you a developer, but they let you validate the AI forecasts, automate what your company's systems don't, and speak fluently with the teams building the tools. The analyst who can audit the model's assumptions is valuable; the analyst who only consumed its output is replaceable.