t7-x

Demand Forecast & Sensing

Configure, audit, and explore the models and data signals that power your demand forecasts.

Forecast Accuracy (WAPE)

96.2%

Within 5% target

Model Bias

+1.8%

Slight over-forecasting

Forecast vs. Actual

-3.8%

LFP sales higher than expected

Manual Adjustments

12

Last 30 days

1. Forecasting Model

Choose the core method used to generate the forecast.

Statistical

e.g., ARIMA, exponential smoothing

Machine Learning

e.g., regression, random forest

AI-Enhanced Hybrid ✅

Patterns, seasonality, and real-time triggers

Tip

Hybrid model is recommended for environments with high variability or external signal influence (EV incentives, market shifts).

2. Forecast Inputs

Select which data sources the forecast engine should use.

Historical Sales Orders

Supplier Confirmations

Production Backlogs

Channel Orders / Retail Sell-Through

Manual Adjustments

External Factors

3. Filters & Grouping

Set the default way demand forecasts should be filtered and grouped.

Group By

SKU

Product Category

Market Segment

Custom

Default Filters

Region

Product Line

Customer Tier

4. Forecast Horizon Settings

Configure default timeframes and calendars.

Default Timeframes

Short-Term (1–3 months)

Medium-Term (3–9 months)

Long-Term (12–24 months)

Rolling (e.g., 90 days)

Calendar

Use Fiscal Calendar

5. Assumptions & Overrides

Fine-tune the forecast logic with custom rules and adjustments.

Weight recent data more heavily

x1.5 weight

Seasonality Toggles

Use trailing 12-month pattern

Use fixed seasonal adjustments

No seasonality

Outlier Handling

Auto-smooth large spikes

Flag for manual review

Manual Override Rules

Quick Actions