Step 02 · Prediction Engine
In progressWhat will happen?
Predict the incremental volume this outlet delivers if it receives a cooler.
An uplift model trained on outlets that received coolers versus matched outlets that did not, reported with held-out accuracy, so the number can be argued with on its merits.
Predicted incremental
+38UC / month
Held-out R² 0.71
Terms on this page
- Uplift model
- A model that estimates the difference an action makes: here, the extra volume a cooler adds.
- Matched outlets
- Outlets without a cooler, chosen to resemble the ones that received one.
- Held-out R²
- How much of the variation in real results the model explains, measured on data it never saw in training. 1.0 is a perfect fit.
How the Prediction Engine gets there
- 01Three inputs: results already surfaced by the Analytics Engine, the underlying database, and the user’s prompt.
- 02Our proprietary agentic structure reads the shape of the data (task type, target, features, volume) and composes a prediction specification.
- 03It selects candidate models, or an ensemble, from a curated catalog, then runs a real ML workflow: preparation, training, validation, held-out testing.
- 04The lifecycle stays prompt-driven: promote a model into production, wire it into a dashboard, schedule re-forecasts and alerts.
In progress. Under active development now: agentic AutoML with calibrated, explainable output for real business prediction targets.
Explore the Prediction EngineWhat changes
The relay today
Request routed to a tech lead, then a data team: gather, clean, build, validate, ship. Weeks or months pass.
With Theovya
A trained, validated model with honest metrics, while the question is still relevant.
Same budget. Different outlets. The delta between a judgement-based allocation and an optimized one is the entire argument for the third engine.