Feature · Future views

Test the future before you bet on it.

One driver-based model, three ways to ask “what’s next?” - forecast the likely path, change a single assumption, or run the full range of what could happen. Every number deterministic, and traceable to its inputs.

Book a demo See how it works
One model, three modes

Choose how much to pin down

ForecastThe most likely path from your history
What-ifChange one assumption, all else equal
SimulationThe full range of what could happen
One model

One engine, three ways to ask “what’s next?”

The same driver-based model underneath - you just choose how much to pin down.

ONE DRIVER-BASED MODEL

Forecast

The most likely path, given your inputs.
You bring
your history

What-if

What happens to the outcome if one assumption changes, all else equal.
You bring
history + a single value

Simulation

Across everything that could happen - which outcomes are likely, and how achievable.
You bring
history + a range
Forecast

The most likely path, from your history.

Project forward with deterministic methods you can inspect - and judge the fit before you trust it.

  • Deterministic methods. Transparent, reproducible projections computed from your data - not a curve an AI drew across a chart.
  • Diagnostics you can read. MAE, RMSE, sMAPE and confidence intervals show you how good the fit really is.
  • Driver-based. The forecast moves with the same inputs that drive the rest of your model.
  • A growing method set. Transparent methods today, with more added over time.
Forecast

The most likely path

You bringYour history and the drivers behind it
MethodDeterministic and inspectable - no black box
You getA projected path with a confidence range
Judge itFit diagnostics you can read before you trust it
What-if

Change one assumption

BaselineYour current plan, left untouched
ChangeFuel +4%
ResultThe outcome recomputes end to end, all else equal
TrackedEvery override is recorded against the baseline
What-if

Change one assumption. See the outcome.

Override any input with a single new value and watch the result move - cleanly, with the baseline still in view.

  • Edit any driver. Set fuel +4%, churn −2%, a price change - and the model recomputes end to end.
  • All else equal. Isolate the effect of one assumption, so you know what caused the change.
  • Every override tracked. See exactly what you changed, against the original baseline.
  • Same model underneath. No separate spreadsheet, no copy that quietly drifts out of sync.
Simulation

The full range of what could happen.

Give an input a range instead of a number, and Fragiliti runs thousands of reproducible scenarios - so you see not just the likely outcome, but how achievable your target is.

  • Ranges, not point guesses. Fuel −1% to +8%, demand within a band - describe the uncertainty you actually face.
  • Monte Carlo, made reproducible. Seeded runs mean the same setup returns the same distribution, every time.
  • Likelihood and achievability. See the spread of outcomes and the odds of hitting your number.
  • Defensible under scrutiny. Every run traces back to the same governed model.
Simulation

The full range of outcomes

You bringA range instead of a single guess
InputsFuel −1% to +8%Demand band
RunsThousands of seeded, reproducible scenarios
You getLikely outcomes and the odds of hitting your target
One model underneath

The future views are only as trustworthy as the model beneath them.

Every projection runs on the same explicit, driver-based model - so you can trace any number back to the inputs that made it, and drop the result straight into a decision.

See data flow & drivers → Turn it into a story →

See the future views on your numbers.

Bring a real decision you’re weighing. We’ll forecast it, flex the assumptions, and run the range - in one session.

Book a demo Request early access