Demand trend and planning scenario
History in black, AI forecast in orange and any context adjustment in blue.
Demonstration environment
Forecast and inventory risk diagnostic.
Upload one product's historical demand and test how a forecasting baseline translates into service and inventory risk.
A forecast readiness check, not a forecasting tool alone.
- What it checks. Your demand history line by line: whether the periods are complete, whether the units stay the same, and what shape the demand actually takes.
- What it tells you. Which lines a forecast can be trusted on, which are waiting on a decision from you, and which no forecasting method will ever fix.
- What it refuses to do. Guess. Where the data is ambiguous it declines to forecast the line and names the answer it needs from you.
- What happens to your file. It is processed for this request and not deliberately retained. The confidential record of a run stays in your browser until you download it.
Use synthetic or anonymised data only. Files are processed for the current request and are not deliberately retained by this application.
Forecast demand
Minimum inventory
Safety stock
Action
Silurian assessment
Projected inventory position
Expected inventory after forecast demand and confirmed inbound supply.
AI management interpretation
AI-generated forecast interpretation. Validate demand assumptions, confirmed supply and operational context before acting.
Forecast back-test
Simple forecasting methods are tested against withheld demand history. The strongest baseline is selected using the lowest WAPE.
| Method | WAPE | MAE | RMSE | Bias |
|---|
Portfolio diagnostic
Prioritise exceptions across multiple SKUs.
Validate demand history first. If the file also contains inventory and supply inputs, the diagnostic can then rank service and working-capital exposure.
Assess data quality to see how much of this portfolio is ready to forecast.
- Analysis date
- Not set
- Frequency
- Not inferred
Run identifiers
- Source
- No file selected
- Run
- Not created
Kept so a run can be reproduced and defended. Nothing here changes a planning decision.
WaitingNot assessed
Input validation
Can this file be used?
Whether the file can be read and trusted at all, and what would have to change if not.
Data quality portrait
Where does context matter?
What the demand history is like, line by line, before anything is forecast.
Demand classification
What kind of demand is this?
What shape each line's demand has, and how much of your volume it carries.
Run data quality to classify the portfolio.
Forecast routing
Which lines may be forecast?
Which lines may be forecast, which may not, and why.
Run data quality to route the portfolio.
Open items
What is waiting on you?
The lines that cannot move until you answer something, biggest first.
Run data quality to list the open items.
Forecast
What can be forecast?
What is ready to model once the open items are settled.
Run provenance
What produced this result?
What produced this result, and what you need to reproduce it.
Glossary
What does that word mean?
Every term this tool uses, in a planner's language.
Loading the glossary.
Silurian data quality portrait
SKUs analysed
Periods covered
Clean volume
Portfolio work surface
| Detail |
|---|
SKUs analysed
Red exceptions
Amber exceptions
Excess risks
AI portfolio interpretation
AI-generated forecast interpretation. Priorities should be reviewed against current supply commitments and operational context.
Prioritised management exceptions
| SKU | Status | Action | Forecast demand | Minimum inventory | First breach | Back-test WAPE |
|---|
Portable run record
Keep, reopen or reproduce a completed run.
The manifest is the proof and is safe to keep anywhere. The bundle is the evidence and belongs to the client.
A run bundle contains client data, including SKU identifiers, metrics, findings and forecast values. Treat it as confidential and do not share it like a manifest.