Interactive Analyze Your Data Lesson 3: run a simple forecast with a range and sort honest forecasting from false confidence, with an AI Coach.
Analyze Your Data · Lesson 3Pro+  ~16 min · Advanced
Pro+

Forecasts & living dashboards.

Reports look backward; forecasts and dashboards look forward — and keep looking. This lesson builds simple, honest forecasts and dashboards that update on their own, so you're watching the trend, not rebuilding the report. Plus how not to over-trust a forecast.

The mental model

A forecast is a range with assumptions, not a promise — and a dashboard is a report that refreshes itself.

AI can project a trend forward and stand up a dashboard that updates as new data arrives. The value is staying current without rebuilding; the trap is treating a projection as a guarantee. A good forecast comes with its assumptions and a range, so you know how much to lean on it.

The reframeA forecast is a conversation starter, not the answer. Use it to ask "what would have to be true?" — not to bet the business on a single number AI drew through your dots.

Step 01 · Build a simple forecast

Press run to project the half-year trend forward — note what an honest forecast includes:

history  6 months: 42 → 59k, ~+40% with a Feb dip
method  simple trend projection, last 6 months
forecast  next month ~62k (range 56–68k)
assumes  no big change in demand or pricing ← the caveat
Forecast promptForecast next [period] from this history. Show the projection as a range, not a single number. State the method and the assumptions it depends on. Tell me what would make this forecast wrong, and how confident I should be.
An honest forecast gives a range, names its assumptions, and says what would break it. A single confident number with no caveats is the one to distrust.

Step 02 · Honest forecast, or false confidence?

Tap each forecast behavior — keeps you honest, or sets you up to be wrong?

Step 03 · Living dashboards

Connect the dashboard to a live source (a sheet, an export) so it refreshes as data arrives — you watch the trend instead of rebuilding the report each week. Start read-only and keep checking the numbers.

An auto-updating dashboard can be confidently, silently wrong. A broken data feed or a changed column quietly corrupts every chart downstream. Spot-check after each refresh, and never let a forecast's tidy line override what you actually know about the business.

Your challenge: forecast and watch a trend

  1. Have AI forecast next period as a range, with assumptions.
  2. Write down what would make the forecast wrong.
  3. Build a dashboard connected to a live source, read-only.
  4. Spot-check it after a refresh before trusting it.

Analyze Your Data — complete

You can turn raw data into answers, the right charts, and honest forecasts and dashboards that update — without writing a formula, and without over-trusting the output. Ready for the next one?

Next build

Automate Your Documents — proposals and reports at scale

🎓
AI Coach
Ask anything about this lesson
Hey! I'm your AI Coach for this lesson. Ask me anything about building forecasts, living dashboards, or not over-trusting a projection. What's on your mind?
Free lesson coaching is limited to 3 questions. Upgrade to Pro for unlimited coaching on every lesson.