CareerConnect
An electronic recruitment platform that lets organisations post jobs, screen and rank applicants with AI, and schedule interviews, while candidates build rich profiles an...
Predict case duration and manage court backlogs proactively
A machine-learning platform that predicts how long court cases will take, from the day of filing and as they progress, and flags high-risk cases early so courts can manage backlogs before they build up.
Case backlogs build up quietly: by the time a case is flagged as delayed, months have already been lost. JaliHaki gives courts an early warning. It predicts how long each case is likely to take and classifies it into a risk tier, so registrars and judicial officers can act early.
JaliHaki uses a three-stage machine learning pipeline trained on 672,974 cleaned case records from Kenyan magistrate courts across 47 counties. Stage 0 identifies cases likely to be disposed of on the day of filing (ROC-AUC 0.91). Stage 1 predicts duration using only filing-time information, and Stage 2 refines the prediction as hearings and adjournments happen.
Predictions come with explanations (SHAP), showing which factors drive each estimate, and every case is placed in an empirically derived Low, Medium, High or Critical risk tier. JaliHaki connects to existing case management systems through a secure prediction API.
Estimate case duration the day a case is filed, using case type, category, court and representation.
Update the estimate as the case progresses, using hearings held, adjournments and first-mention delay.
Identify cases likely to be resolved on the filing day, such as pleas and withdrawals.
Low, Medium, High and Critical tiers derived from the data, to focus attention where it matters.
SHAP explanations show which factors drive each prediction, supporting transparent decisions.
A containerised REST API that integrates with existing case tracking and management systems.
JaliHaki adapts to organisations of every size, from a single office to a national network.
Anything else you would like to know? Ask us during your demo.
A product specialist will walk you through JaliHaki using scenarios from your own organisation, and answer your questions on pricing, set-up and support.
Tell us a little about your organisation and we will tailor the demo to you. It takes about 30–45 minutes, online or at your office.
An electronic recruitment platform that lets organisations post jobs, screen and rank applicants with AI, and schedule interviews, while candidates build rich profiles an...
An offline-first shopping management app that helps households and businesses plan shopping lists, see what a trip will cost before reaching the till, stay within budget...
Rental property management software for Kenyan landlords and property managers: automatic invoicing, M-Pesa rent collection and reconciliation, arrears follow-up, mainten...