Guidewire Kaleidoscope Documentation

Build Better Claims Models with Industry-Wide Data

Kaleidoscope enables predictive modelling teams to train models on Guidewire Industry Intel data alongside their internal data. Where third-party data adds new attributes to the data you already have, Kaleidoscope adds multi-carrier data volume to strengthen models for thin segments, new lines of business, and low-frequency events. Federated learning moves the analytics and model training to where each dataset lives, so record-level data is never exposed between your organization and Guidewire.

Getting Started

Sign in, connect your data, and run your first session.

getting-started.html

Federated Model Training

Learn about vertical and horizontal federated learning models.

train-overview-vfl.html

Exploratory Data Analysis

Explore different techniques for data analysis.

analyze-overview.html

Data Preprocessing & Feature Engineering

Learn how to use the transform session and other tools.

preprocessing.html

Release Notes

Updated for each release.

release-notes.html