Forecasts & predictions
See what is coming, and how sure to be.
Forecast any measure in the graph: several model families compete on back-tests and the best one wins, with a range taken from the model itself. Predict who is at risk or ready with models trained on your own history, each with a model card.
Attrition · Engineering, projected 6 months
Forecasts & predictions, part by part.
- 01
Ranges you can trust
On AutoGluon forecasts, 80–99% ranges are the chosen model's own quantiles, never stretched, with how often each held on back-tests. A history too short for back-tests uses classic methods, and the card says so.
- 02
Can I trust this?
Every model fitted is ranked on back-tests, and the forecast gets a grade with its reasons in words. Not convinced? Fit it again with a stronger setting.
- 03
Models you can read
Attrition risk, promotion and readiness models come with a model card: how it did, its drivers and how it errs — and a fairness check on yes/no predictions, by gender and age band for attrition risk.
- 04
What-ifs, labelled
Change an input and see the result again: exact on an answer, projected on a forecast, estimated on a model. Or find the change that reaches a goal.
Pick, Fit, Check, Act.
- Step 1
Pick
Any measure IntRpret can count over time.
- Step 2
Fit
Model families compete on back-tests.
- Step 3
Check
The range, the grade and its reasons.
- Step 4
Act
Ask what-if, and keep the scenario you chose.
What competes for each forecast
- Statistical
- Seasonal naive, exponential smoothing (ETS) and Theta.
- Machine learning
- Gradient-boosted trees (LightGBM).
- Pretrained
- Chronos-2 and Toto-2 time-series foundation models.
- Ensemble
- A weighted blend of the best.
- Predictions
- LightGBM, CatBoost, XGBoost, random forests, nearest neighbours, neural networks and linear models, tested on held-out records.
Questions it answers, in your words.
- “Forecast our headcount for the next 12 months.”
- “Who in Customer Support is most likely to leave this year, and why?”
- “What happens to attrition if we close the pay gap in grade 5?”
On the same model and graph
- Decision IntelligenceKPIs built by an agent and watched every month, with alerts on the change and the cause behind it.
- Workforce IntelligencePlan the workforce you will need, and who could step into its critical roles.
- Analytics & reportingCharts chosen to suit each result, with the working and the downloads behind them.
Your next people decision, with the why attached.
Request a demo. We'll email you a two-minute walkthrough today, then set up a time that suits you.