Choosing an analytics engine
KEZEL Insights Studio includes regression, classification and time-series models. The method should follow the decision you need to make.
Forecasting demand and ranking review cases require different models. KEZEL Insights Studio, introduced at Bengaluru Tech Summit 2025, groups its engines by those analytical tasks.
Regression estimates a relationship, such as campaign spend against weekly volume. The fitted relationship can support a simulation, but it does not establish that changing one variable causes the other to change. An analyst needs to check that assumption before using the estimate.
Classification produces ranked scores with reasons attached. Insights Studio evaluates eight model candidates, with and without outlier correction. A reviewer can inspect the reason for a case's position in the queue, then make the decision.
Read the uncertainty alongside the forecast
The time-series engine uses Prophet and ARIMA model families. Its projection includes a confidence band. A wide band matters when deciding how much stock to order or when capacity will be needed.
These methods supply estimates and rankings. Interpreting a cause, choosing an intervention and checking the model's suitability still require context from the team using it.
All three engines run beside the system of record, with results read under policy. Compare the sample engine outputs against the question you want a pilot to answer.
written by the KEZEL team at DBTEZ · all articles · book a demo