Working capital is where financial strategy meets operational reality. CFOs and controllers know the symptoms well: cash trapped in inventory that isn’t moving, receivables stretching past terms, payables paid early out of habit rather than policy, and a cash conversion cycle that quietly lengthens quarter after quarter. The frustrating part is not the lack of data. ERP systems, treasury platforms, and bank feeds generate more numbers than any finance team can realistically digest. The problem is turning that raw data into decisions — which customers to chase, which SKUs to liquidate, which payment terms to renegotiate, and which levers will free the most cash with the least commercial damage.
This is exactly where ChatGPT becomes useful. Not as a system of record, but as an analytical partner that can restructure messy working capital data, model scenarios, draft negotiation positions, and pressure-test the assumptions behind your cash forecast. It compresses hours of spreadsheet wrangling into minutes of structured reasoning, and it does so in plain language your whole finance team can follow.
The key is prompting with discipline. Vague requests produce vague output. The prompts below are built around a structured “anatomy” format — task, context files, reference patterns, success brief, and an explicit instruction to plan before executing. That structure is what separates a generic chatbot answer from a board-ready working capital analysis.
Why structure beats improvisation in finance prompts
When you ask ChatGPT to “help with working capital,” you get textbook definitions of DSO, DPO, and DIO. When you give it your actual aging buckets, your terms structure, and a clear definition of success, you get something closer to a junior analyst’s first draft — one you can refine rather than rewrite. The two prompts below follow that logic. The first focuses on diagnosing and prioritizing cash release opportunities. The second focuses on building a 13-week cash flow forecast narrative you can present to leadership.
First, read these files completely before responding:
aging_report.csv — AR aging buckets by customer, 0-30/31-60/61-90/90+ days, with invoice values
ap_ledger.csv — AP balances by vendor with payment terms and actual payment dates
inventory_summary.csv — SKU-level inventory value, turns, days on hand, and slow-mover flags
working_capital_policy.md — our current DSO, DPO, and DIO targets and credit policy rules
Here is a reference for what I want to achieve:
[Upload a prior working capital review deck or memo as markdown]
Here’s what makes this reference work:
It leads with a one-page cash impact summary, quantifies each lever in currency, ranks actions by effort versus impact, and assigns an owner and deadline to every recommendation.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: Structured diagnostic memo, 800-1000 words, with a ranked action table
Recipient’s reaction: The CFO should think “this is actionable this quarter” and forward it to the controller without edits
Does NOT sound like: A textbook explanation of working capital ratios or a generic consulting framework
Success means: At least five quantified cash release actions, each with an estimated value and an owner
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Give me your execution plan (5 steps max) before you begin.
From diagnosis to forecast
Once you know where the cash is trapped, the next question is timing. A working capital diagnostic tells you what is possible; a rolling cash forecast tells you when it becomes real. Most finance teams underestimate how much of the forecast depends on behavioral assumptions — will the top ten customers actually pay within terms this quarter, or will the two chronic late payers slip again? ChatGPT is good at surfacing those assumptions and forcing you to state them explicitly, which is precisely what makes a forecast defensible in front of a board or a lender.
The second prompt builds that forecast narrative. It takes your historical collection patterns, your payment run schedule, and your inventory replenishment cycle, and produces a 13-week view with clearly labeled assumptions. The value is not the number itself — your treasury system owns that — but the reasoning trail behind it.
First, read these files completely before responding:
collections_history.csv — 12 months of invoice dates, due dates, and actual payment dates by customer
payment_schedule.csv — scheduled vendor payments, payroll, tax, and debt service for the next 13 weeks
inventory_replenishment.csv — planned purchase orders and lead times by supplier
cash_position.md — current cash, revolver availability, and minimum operating cash threshold
Here is a reference for what I want to achieve:
[Upload a prior board cash forecast memo as markdown]
Here’s what makes this reference work:
It states every assumption in one line, separates committed from probable cash flows, highlights the lowest projected cash week, and includes a sensitivity note on the two largest variables.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: 13-week forecast narrative, 900-1200 words, with a weekly summary table and assumption log
Recipient’s reaction: The board should see disciplined thinking, not optimism, and trust the downside case
Does NOT sound like: A single-point forecast presented as certainty, or a spreadsheet dump without interpretation
Success means: Every weekly figure traceable to a stated assumption, with the lowest cash week identified and a mitigation option attached
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Give me your execution plan (5 steps max) before you begin.
A practical tip before you start: run the diagnostic prompt first, then feed its ranked action table into the forecast prompt as an additional context file. The two outputs compound — the diagnostic tells you which levers matter, and the forecast tells you when their cash impact lands. Keep your context file current, because ChatGPT will hold you to the standards you define in it. If your working capital policy changes mid-quarter, update the file and rerun.
Start with one business unit or one region rather than the whole enterprise. A narrow scope produces sharper output, and a successful pilot gives you the credibility to scale the approach across the finance organization.
Published on 30 September 2026 on growwithgpt.com