A forecast is useful when someone can change an assumption and understand what happens next. It is much less useful when the numbers arrive from a model that hides the source actuals, mixes inputs with formulas, or produces scenarios no one asked for.

Strawberry can work in the existing spreadsheet while checking approved operating context in the browser and connected apps. Your companion can keep source versions and assumptions visible, update the calculations, and carry accepted performance drivers into the forward view.

Start with the decision

A hiring decision, runway discussion, annual budget, and rolling forecast need different horizons and drivers. Define the company, time grain, currency, audience, and decision before changing the model. Preserve a reliable existing model rather than replacing it with a generic template.

Want to try it?

Ask your Strawberry companion: “Build or update our forecast from accepted actuals and operating drivers, make the material assumptions clear, and test the scenarios this decision needs.”

Skill

Build or update a financial forecast

Build or update the model around the decision, accepted actuals, and operating drivers that matter.

Make the model inspectable

LayerWhat to keep clear
ActualsAccepted source, version, period, currency, and as-of date.
Operating inputsContracts, pipeline, pricing, usage, headcount, vendor commitments, and cash evidence.
AssumptionsOwner, rationale, date, and what would cause the assumption to change.
Calculations and outputsFormulas, dependencies, scenario switches, and the decisions the output informs.

Historical patterns can inform an assumption, but they do not prove that it will continue. Keep that choice visible rather than burying it inside a formula.

Use scenarios to answer real questions

Build the accepted base view first. Add an upside, downside, or sensitivity only when it changes a real decision. A useful scenario changes a few meaningful drivers and explains what they represent; it is not a second spreadsheet full of arbitrary percentages.

  • For runway, test collections, hiring, revenue timing, one-time costs, and financing only where they are material.
  • For hiring, show the start date, compensation, related costs, and operating effect.
  • For revenue, separate volume, price, timing, churn, and collection assumptions when the distinction matters.

Validate before sharing the view

Reconcile the opening actuals and period roll-forward. Check totals, signs, formulas, dates, units, cash movements, linked schedules, and scenario switches. Investigate implausible jumps and disclose stale inputs, missing drivers, or manual overrides.

Deliver the model with a concise assumption log, the base view, decision-relevant scenarios, cash or runway implications, and the questions that remain. Once the team trusts the model logic, save the method as a team skill so later updates use the same drivers and checks. A Routine can prepare a recurring draft and stop when the actuals, model version, or operating assumptions no longer match the accepted setup.

Overwriting an approved budget, publishing the forecast, or changing an operating plan remains a separate decision.

Skill

Review financial performance

Compare future actuals with the accepted forecast and explain the material drivers.

How to organize invoices and receipts with AIPrepare traceable records and resolve gaps before they become model inputs. How to review financial performance with AICompare future actuals with the accepted forecast and explain the important movements. How to prepare for a fundraise with AICarry the accepted financial view into the wider investor story and fundraising process.
Official Strawberry skill
Copy skill file Download skill file
name
build-a-financial-forecast
description
Build or update an assumption-led financial forecast with accepted actuals, operating drivers, scenarios, cash, and runway where relevant. Use for startup budgets, rolling forecasts, cash planning, hiring or revenue scenarios, and decision-specific financial models without changing approved source records.

Build a Financial Forecast

Build a forecast whose assumptions and calculations can be inspected and changed. Preserve a reliable existing model rather than replacing it with a generic template.

1. Define the decision

Establish the company or entity, horizon, time grain, currency, intended decision, audience, and required outputs. Identify whether the user needs a budget, rolling forecast, cash view, runway model, or one decision-specific scenario.

Confirm the accepted actuals and current source model. Separate assumptions the user owns from inputs drawn from operating systems and questions that remain open.

2. Build from accepted drivers

Use approved actuals, contracts, pipeline, headcount plans, pricing, usage, vendor commitments, cash balances, and other evidence that genuinely drives the model. Preserve sources, versions, dates, and owners for material inputs.

Keep inputs, assumptions, calculations, and outputs distinct. Make units, signs, dates, currencies, and dependencies explicit. Use historical patterns as context, not proof that they will continue.

3. Test the uncertainty that matters

Build the accepted base view first. Add an upside, downside, sensitivity, or decision scenario only when it answers a real question. Change the few assumptions that define each scenario and explain their operating meaning.

When cash matters, connect profit, working capital, funding, and cash timing carefully. Show how runway changes with collections, hiring, revenue, one-time costs, or financing rather than presenting one date as certain.

4. Validate and deliver the model

Reconcile the opening actuals and period roll-forwards. Check totals, signs, formulas, dates, units, scenario switches, cash movements, and linked schedules. Investigate implausible jumps and disclose missing drivers, stale inputs, and manual overrides.

Provide the updated model or reviewable draft with:

  • scope, version, source actuals, and as-of date;
  • a concise assumption log with owner and evidence;
  • the base view and decision-relevant scenarios;
  • cash and runway implications when in scope;
  • the largest changes from the prior view; and
  • sensitivities, gaps, decisions, and next update points.

Use strawberry/finance/review-financial-performance when the next question is how actual results compare with this forecast.

5. Keep the forecast controlled

Preserve accepted sources, definitions, driver logic, scenario rules, and review preferences after the user confirms them. A forecast Routine may refresh named inputs and prepare a proposed update, but it should stop when a source version, formula, company, assumption owner, or business condition falls outside the accepted model.

Do not overwrite an approved budget, publish a forecast, change an operating plan, or write back to source systems. Treat those as separate actions outside this release.