Cash divided by last month's burn says the company has fourteen months. The hiring plan says nine. The annual software renewals and tax bill say less. The fundraising plan assumes a round closes before any of those versions become uncomfortable. Only one cash balance exists; the rest are choices and assumptions.
Startup cash discipline begins with three plain facts: cash in the bank, money coming in and money going out. The default-alive test asks whether that trajectory reaches sustainability. Unit economics explains which growth choices improve or consume the path. The synthesis stays simple on purpose: precision cannot rescue dishonest inputs.
Turn cash into decision time
Runway is time to learn, decide and still change course.
01 · Cash
Observe balance
Use available company cash, not promised money.
02 · Burn
Normalise flow
Separate recurring spend and true one-offs.
03 · Scenario
Add choices
Include hires, renewals, revenue and downside.
04 · Decision
Set the date
Know when to raise, cut, sell or reach breakeven.
Keep three runway views
The baseline uses current cash, recent average inflows and recurring outflows. The operating plan adds commitments the company intends to make. The downside case delays revenue, includes lumpy obligations and assumes fundraising takes longer or does not happen. Each view needs a date, inputs and owner, not a single coloured cell.
- Baseline: what happens if the company continues as it is?
- Plan: what happens after approved hiring, spend and expected revenue?
- Downside: what happens when timing slips and the next round is unavailable?
Treat runway as a budget for learning
A twelve-month runway can be adequate for a company close to breakeven and dangerous for one needing six months to learn, six to raise and three to close. Work backwards from the next material milestone and the time required to respond if it fails. Paul Graham's default alive or default dead test is useful because it forces the company to separate survival on its trajectory from hope of another cheque.
Let AI watch the assumptions, not move the money
An agent can reconcile source freshness, recompute scenarios and flag when an assumption diverges from actuals. It should state which bank, ledger period and plan revision support the answer. It should not treat forecast revenue as cash, initiate payments or revise the accepted budget without review.
Train the agent on a finance source hierarchy
Finance work needs a strict order of authority. The agent should know which source establishes cash, which ledger period is closed, which plan is accepted and which numbers are still assumptions. Without that hierarchy, fluent scenario work only makes uncertainty harder to see.
- Evidence pack: dated bank balances, reconciled ledger periods, billing and payroll evidence, signed contracts and known renewals, the accepted hiring and spend plan, forecast assumptions, currency policy and the last approved runway scenario.
- Working brief: reconcile source freshness first, then produce baseline, operating plan and downside views from the same observed cash. Label every input as observed, committed or assumed and show the sensitivity that moves a decision date.
- Permitted action: recalculate scenarios, flag stale or broken sources, prepare the weekly variance note and draft an editable plan revision when an agreed threshold is crossed.
- Review gate: people accept budgets, forecasts, hiring and spending decisions. The agent must not initiate a payment, edit the ledger, count unsigned revenue as cash or make accounting and tax judgements.
- Learning record: retain the accepted metric definitions, scenario revision, founder edits and later actuals. Use forecast misses to correct the relevant assumption rule rather than making every future estimate uniformly conservative.
The agent improves through forecast-versus-actual evidence and reviewed definitions, not through confidence. Each run becomes easier to audit and more useful because the same source hierarchy survives the conversation.
A useful next move
Build baseline, plan and downside runway from the same cash source. Mark three dates: start fundraising, switch to the downside plan and reach the minimum cash reserve. Review weekly whenever any decision is less than a quarter away.
Continue with how to make startup metrics useful for decisions, or read when to start a fundraise.