Agent use case / Finance

A 13-week cash forecast, rebuilt every Monday with its assumptions showing

ZeroTwo starts from your reconciled opening cash, dates every receipt and payment it can find, runs base, downside and upside cases, and posts a brief to your leadership channel. Anything it could not date is listed as a warning. A finance owner approves the result.

Set up your Monday forecast See an illustrative workbook

#finance-leadership

Illustrative layout

Monday cash brief, week 1 of 13

Opening cash: $1,224K
Week 13, base: $486K
Week 13, downside: $267K
Week 13, upside: $602K

Alert: downside ends below $300K in week 13.

Warnings: 2 commitments have no date. Not approved until reviewed.

Synthetic numbers for a fictional company, to show the layout. Not a customer result.

Each Monday

Four things land in the channel

The point is not a prettier spreadsheet. It is a forecast you can interrogate: every number traces to an input, and every gap is named.

1
Reconciled opening cash: Bank balance adjusted for items in flight.; 13
Weeks, dated line by line: Receipts, payroll, vendors, tax and debt.; 3
Scenarios with stated assumptions: Base, downside and upside side by side.; 0
Silent assumptions: Missing sources appear as warnings.

Illustrative workbook

What the 13-week table looks like

A synthetic example for a fictional company, with round numbers chosen to show the mechanics. It is a template for what you should expect to review, not a measured outcome. Amounts are in thousands of US dollars.

Opening cash reconciliation

Bank statement balance: $1,240K
Outstanding payments: -$38K
Deposits in transit: $22K
Reconciled opening cash: $1,224K

Every later week is built from this number. If it is wrong by $50K, every week is wrong by $50K, so a person checks it first.

Illustrative 13-week base-case cash table, in thousands of US dollars: receipts, payroll, vendors and cards, tax and debt, net change and ending cash by week.
Week	Receipts	Payroll	Vendors and cards	Tax and debt	Net	Ending cash
1	$120K	-	-$46K	-	$74K	$1,298K
2	$104K	-$212K	-$39K	-	-$147K	$1,151K
3	$98K	-	-$44K	-	$54K	$1,205K
4	$126K	-$212K	-$138K	-	-$224K	$981K
5	$112K	-	-$41K	-	$71K	$1,052K
6	$101K	-$212K	-$37K	-$85K	-$233K	$819K
7	$118K	-	-$43K	-	$75K	$894K
8	$109K	-$212K	-$141K	-	-$244K	$650K
9	$122K	-	-$40K	-	$82K	$732K
10	$99K	-$212K	-$38K	-	-$151K	$581K
11	$115K	-	-$45K	-	$70K	$651K
12	$106K	-$212K	-$44K	-	-$150K	$501K
13	$124K	-	-$139K	-	-$15K	$486K

Base
$486K
ending cash, week 13
Receipts arrive as scheduled. All commitments land on their dates.
Lowest week: 13, at $486K. No alert.
Downside
$267K
ending cash, week 13
Receipts run 15% below base (churn or slow collections). Outflows unchanged.
Breaches the $300K alert in week 13.
Upside
$602K
ending cash, week 13
Receipts run 8% above base (pipeline converts). Outflows unchanged.
Lowest week: 13, at $602K. No alert.

Coverage

What goes in, and what each source cannot see

A forecast is only as complete as its inputs. Connect what your workspace supports, and paste or export the rest from a sheet or CSV. See the connectors list for what can be connected.

Input sources for the forecast: what each supplies and what it cannot see.
Source	What it supplies	Gap to watch
Bank statement	Opening cash, reconciled to your books	You provide this. Without it the whole table is shifted by the unreconciled difference.
Stripe	Expected receipts and payout timing	Does not see invoices paid by bank transfer or outside Stripe.
Ramp	Card, bill and subscription commitments	Does not see vendors paid from other cards, accounts or by hand.
Payroll schedule	Pay dates, amounts, employer taxes, new hires and exits	Read from a connected system or a sheet you keep. Not inferred from bank lines.
Tax, debt and one-offs	Quarterly taxes, loan payments, annual contracts, bonuses	Manual list. The most common source of a surprise.

Timing and risk

Dates cash moves, not dates invoices say

Two choices decide whether a forecast earns trust.

Receipts and payments sit in the week they clear

A Stripe charge is not cash until the payout lands, and a Ramp bill is not cash until the statement is paid. The forecast uses payout schedules, statement dates and vendor terms you supply, and shows any item with no known date as a warning.

Overdue receivables are an assumption, not a fact. When collection delay is the biggest swing in your downside case, pair this workflow with the invoice collections follow-up agent, which works specific overdue invoices.

Where renewal-risk flags come from

A renewal is flagged as at risk only against a usage or health signal you connect or paste, such as a usage export. With no such source, the forecast carries your churn assumption and does not claim to know which accounts will leave.

Beyond week 13

Runway is an extrapolation, so it is labelled one

The 13-week table is the forecast. Months of runway is a separate division: ending cash divided by the average weekly net outflow of recent weeks. It assumes the recent pattern repeats, which is exactly what a one-time AWS bill or a new hire breaks.

On the illustrative workbook, base-case ending cash of $486K divided by the weeks 6 to 13 average net outflow of $71K per week gives about 6 more weeks. That is a rough pointer for a conversation, not a date.

The Monday run

Five steps, same order every week

Set the schedule and channel in the prompt. Time of day is yours to choose, for example early Monday before standup, through scheduled tasks.

Step 1 / Stripe
Reads expected inflows and the payout schedule
Recurring charges, open invoices and the payout calendar become dated receipts. Renewals that look at risk are flagged only when you have connected or pasted a usage source to judge them by.
Step 2 / Ramp
Reads committed outflows and upcoming contracts
Locked-in subscriptions, approved bills and statement dates go in as dated payments. Variable spend is projected from a trailing average you can see and change.
Step 3 / Google Sheets
Adds what no feed knows about
Payroll dates, tax and debt payments, and off-system commitments come from a sheet you maintain. Anything missing is listed as a warning, not silently assumed to be zero.
Step 4 / Google
Builds the 13-week table and three scenarios
One week-by-week table from reconciled opening cash, rerun with the downside and upside assumptions your finance owner has set. Each scenario states its assumptions next to its result.
Step 5 / Slack
Posts the Monday brief and fires threshold alerts
A one-screen summary goes to the channel you choose: ending cash, change since last week, the lowest week per scenario, and any alert that was breached.

Before anyone relies on it

Missing-source warnings and a human sign-off

The forecast is a draft until a finance owner approves it. Check these first.

A bank account or card that is not connected or listed is invisible to the forecast.
Payroll taxes and benefits are easy to leave out when only net pay is entered.
Receipts paid outside Stripe are missing unless you add them.
Annual contracts and quarterly taxes only appear if someone lists them.
Foreign-currency balances need a stated exchange rate, or they are guesses.

Keep a forecast-versus-actual log

Each Monday, compare last week's forecast to what the bank shows, note the largest misses and their cause, and feed corrections into the next run. A few months of that log tells you which inputs to trust and how wide to make the downside. ZeroTwo can draft the comparison, and you decide what is acceptable.

Get started

Paste this, then edit it to match your books

Replace the figures with your own thresholds and add the sources you actually use.

Every Monday, rebuild a 13-week cash forecast. Start from the opening cash I paste, add expected Stripe receipts, Ramp commitments and the payroll sheet, and list anything with no date as a warning. Run base, downside (receipts 15% lower) and upside (8% higher). Post the lowest week per scenario to #finance-leadership and alert if any week of the downside ends below $300,000.

Sample prompt. Amounts and channel names are placeholders.

Keep it in its lane

FAQ

Questions finance teams ask first

How does ZeroTwo handle variable revenue like usage-based billing?

It projects usage-based receipts from the historical billing records you connect or paste, using a trailing average and the trend you tell it to apply. The downside and upside scenarios bracket that projection. It cannot see seasonality it was not given data for, so say so in the prompt if your revenue is seasonal.

Can ZeroTwo include payroll?

Yes, as a fixed schedule. Payroll can come from a connected system or from a Google Sheet that lists pay dates and amounts. New hires and departures only change the forecast when you flag them, and employer taxes only appear when you include them.

How far out can the forecast extend?

Thirteen weeks is the default because receipts and payments can still be dated with some confidence. You can ask for a longer horizon, but treat anything beyond week 13 as a scenario rather than a forecast, and keep it separate from the 13-week table.

Does ZeroTwo account for payment timing?

It places each flow in the week cash is expected to move, using payout schedules, statement dates and vendor terms you supply or that the connected account exposes, rather than assuming everything lands on the invoice date. Items with no known date are listed as warnings.

Can I set alerts for cash thresholds?

Yes. State the rule in the prompt, for example notify the channel if projected cash drops below a set amount in any week of the downside scenario. The check runs each time the forecast is rebuilt, and the Slack post names the week and the scenario that breached it.

Does this replace a finance owner?

No. The forecast is a draft until a named finance owner reviews the opening cash reconciliation, the assumptions and the warnings. ZeroTwo can rebuild the table every week; deciding what the numbers mean for hiring, spend or fundraising stays with you.

Build your first Monday forecast on the Free plan

Free starts with 100 credits and needs no card. Compare plans on the pricing page.

Start a cash forecast