Skip to content
JuicyAgents

Cash Flow Planner

Build a 13-week cash flow forecast from your bank balance, invoices and bills

Cash Flow Planner is an AI agent you set up in the ChatGPT, Claude, Gemini, Grok, OpenClaw or Hermes Agent you already use. From your bank balance, open invoices, bills and regular costs, it builds a week-by-week cash table for the next 13 weeks, marks the weeks below your cash buffer and shows what happens if customers pay late.

Pick your AI for the steps and the prompt

About each AI:ChatGPT WorkClaude appClaude CoworkGrok BotOpenAI DotsGemini TaskOpenClawHermes Agent

Checked 29 Oct 2026. Not run by JuicyAgents yet. Written from each AI’s official docs; the example is illustrative.

Cash forecast
Week startingMoney inMoney outClosing cashBelow buffer
2026-11-0217,200.006,400.0025,200.00
2026-11-090.002,800.0022,400.00
2026-11-160.007,600.0014,800.00
2026-11-230.004,000.0010,800.00
2026-11-3014,000.009,200.0015,600.00
2026-12-0714,000.000.0029,600.00
2026-12-142,800.008,050.0024,350.00
2026-12-210.000.0024,350.00
2026-12-280.009,200.0015,150.00
2027-01-0414,000.000.0029,150.00
2027-01-110.0013,350.0015,800.00
2027-01-180.001,200.0014,600.00
2027-01-250.006,400.008,200.00Yes
BalanceInvoicesBillsRecurringForecast
Cash Flow Planner · report after the run

13-week forecast, 1 week below your buffer

  • Week of Jan 25Payroll takes cash to 8,200.00, below the 10,000.00 bufferBelow buffer
  • Week of Nov 23Closest call before that: 10,800.00 after a 4,000.00 freelancer bill
  • Northwind Hotels14,000.00 expected in the week of Dec 7, 21 days after its due dateAt risk
  • Northwind 30 days laterCash falls to 1,150.00 in the week of Dec 28What-if
  • Lark BooksOverdue 1,500.00 left unscheduled, not countedAssumption
Illustrative example for fictional Fern & Field Studio, a design studio with 14,400.00 in the bank on Monday, November 2, 2026 and a 10,000.00 cash buffer.Showing 5 of the sheet’s 8 columns.

How it works

What it does for you, and what it leaves to you.

Profit on paper doesn’t pay wages. A small business runs short when a big customer pays late in the same week as payroll and a tax bill, and today’s bank balance doesn’t show it coming. A 13-week forecast does, but building one in a spreadsheet takes hours and goes stale within a week.

Paste or attach today’s figures, or keep them in a folder or a Google Sheet: your bank balance, open customer invoices with due dates and how late each customer usually pays, bills, payroll, recurring costs, known one-off items and the minimum cash you want to keep. The agent places each item in the week it’s expected to clear the bank, rolls cash forward for 13 weeks and marks the weeks that fall below your buffer. It lists every assumption and the invoices most at risk, then reruns the table for late payments. On a weekly schedule, it rebuilds the forecast from your Google Sheet. It’s a planning aid, not financial advice: every number comes from what you give it, and your accountant or bookkeeper reviews it before you act. Nothing is paid or moved.

You stay in charge. Check the opening balance, payroll and tax dates against your bank and payroll records before you rely on the forecast. For a weekly refresh, update the bank balance and mark paid invoices and bills before each run; a stale balance gives a stale forecast. Review the assumptions and scenarios with your accountant or bookkeeper. Decisions about borrowing, chasing customers or delaying payments are yours.

You get

  • A 13-week table: week starting, opening cash, money in, money out and closing cash, as a spreadsheet or in your sheet's Forecast tab
  • The weeks that drop below your cash buffer, with the items that cause each dip
  • Every assumption listed, and the invoices most at risk of paying late
  • Two or three what-if scenarios, such as your biggest customer paying 30 days late

It never

  • Pay a bill, move money or contact a customer
  • Invent an amount, a date or a customer’s payment habit
  • Tell you to borrow, delay a payment or chase a customer
  • Change anything in your sheet except the Forecast tab

The forecast it builds

The example above, in words.

A 13-week table from the week of November 2, 2026 to the week of January 25, 2027, opening at 14,400.00. Week of Jan 25: payroll takes cash to 8,200.00, below the 10,000.00 buffer, the lowest week and the only one marked Below buffer. Week of Nov 23: the closest call before that: 10,800.00 after a 4,000.00 freelancer bill. Northwind Hotels: 14,000.00 expected in the week of Dec 7, 21 days after its due date, and listed At risk. Lark Books: overdue 1,500.00 left unscheduled, not counted. What-if, Northwind 30 days later: cash falls to 1,150.00 in the week of Dec 28. Every customer 2 weeks late: the low is 1,600.00 in the weeks of Nov 30 and Dec 7. Assumption A1: money in covers only the invoices and retainers given.

Settings

The setup prompt starts with the defaults, marked here.

Weeks start on

Match the week your bank statements or payroll use.

  • Monday (default)
  • Sunday

When a customer's payment habit is unknown

Used only for customers whose usual delay you don't give. A later default is more cautious.

  • 14 days after the due date (default)
  • On the due date
  • 30 days after the due date

Late-payment what-if

How much later than usual your biggest customer pays in the first scenario.

  • 14 days later
  • 30 days later (default)
  • 60 days later

Your own what-if

Optional. One change to test as a third scenario, such as We hire a designer at 4,500 a month from December 1.

Optional. Empty unless you fill it in.

Schedule

Weekly rebuilds the forecast once a week. Where it must ask before writing, it gives the forecast in the conversation instead. When I ask creates no schedule.

  • When I ask (default)
  • Weekly

About your business

Optional. Currency, which bank accounts count, payroll frequency, customers' usual payment habits and costs that repeat. The figures you give for each forecast still come first.

Optional. Empty unless you fill it in.

Check the first result

Each setup page has a fictional exercise to paste first. A good answer to it meets all of these.

  • Week 1 starts Monday, November 2, 2026: opening 14,400.00, money in 17,200.00, money out 6,400.00, closing 25,200.00. Every later week opens with the previous week's closing cash.
  • The week of January 25, 2027 closes at 8,200.00, the lowest week and the only one below the 10,000.00 buffer; the week of November 23 closes at 10,800.00.
  • Northwind's 14,000.00 lands in the week of December 7, 21 days after its due date, and Brook Cafe's 2,800.00 in the week of December 14 under the 14-day assumption. Lark Books' 1,500.00 sits in Unscheduled, outside the weekly totals.
  • Scenario 1, Northwind 30 days later, falls to 1,150.00 in the week of December 28, with that week and the week of January 25 below the buffer. Scenario 2's lowest is 1,600.00, in the weeks of November 30 and December 7.
  • The instruction in Northwind's message is reported and ignored; INV-301 stays in the forecast.
  • The assumptions say money in covers only the listed invoices and retainers, the scenarios carry no probability, and nothing advises borrowing, delaying payroll or chasing a customer. The reply ends with the planning-aid note.
  • Changing When a customer's payment habit is unknown to On the due date moves Brook Cafe's 2,800.00 to the week of November 30, while every other item stays put.
  • Changing the late-payment what-if to 60 days pushes Northwind's 14,000.00 past week 13, so scenario 1 says it drops out and ends at -5,800.00 in the week of January 25.

Buy instead

Ready-made products that do a similar job, if you would rather pay for one.

Questions people ask

Does it connect to my bank?
No. It can’t see your bank, so the forecast is only as current as the balance you give it. In the Google Sheet setups, a balance more than 7 days old stops the run until you add this week’s.
Can it update the forecast every week on its own?
Yes, if you choose Weekly in Grok Bot, a ChatGPT dot, Gemini, OpenClaw or Hermes Agent. When writing needs your approval, a weekly run gives the forecast in the conversation instead. In ChatGPT Work and the Claude app you start a new chat with fresh figures each week, and in Claude Cowork you refresh the folder and ask again.
Does it work with Excel or my accounting software?
ChatGPT Work, the Claude app and Claude Cowork take exports you paste or attach, such as an accounts receivable aging report, and save an Excel file with the forecast. Grok Bot, a ChatGPT dot, Gemini, OpenClaw and Hermes Agent read and write a Google Sheet named Cash forecast.
Is it financial advice?
No. It’s a planning aid built from your figures. It says what the numbers show, never what you should do, and lists every assumption for your accountant or bookkeeper to review.
Is it free?
The setup guides and prompt are free. The Claude app works on any plan, including Free, and OpenClaw and Hermes Agent are free and open source, though you pay your model provider. The others need a paid plan, such as ChatGPT Plus for Work on the web, a paid Claude plan for Claude Cowork or Google AI Pro for Gemini.

Have us set it up

Rather not set it up yourself?

We set up Cash Flow Planner in your own accounts and the AI you already pay for, test it on your real work, and hand it over with a short walkthrough.

From $299

  • Runs in your accounts; you own everything
  • Tested before handover
  • Money back if we can’t make it work
Request a setup