s1r3n
AI & Data Science UDOMJuaHaki
Problem Statement
Navigating Tanzania's justice system is confusing people often don't know which court, tribunal, or legal-aid body their case actually belongs to. When a matter lands in the wrong forum, it leads to re-filings, delays, extra costs, and more backlog for the courts. JuaHaki tackles this by using machine learning to help both citizens and justice sector staff figure out the right next step for a legal matter, acting as a decision support tool rather than a chatbot or a replacement for legal professionals.
Proposed Solution
JuaHaki a machine learning-based decision support system that takes basic details about a legal matter and recommends the right justice pathway. It works in three layers: a classification model predicts the likely forum (Primary Court, Ward Tribunal, DLHT, CMA, etc.), that prediction gets checked against actual jurisdiction rules and statutory thresholds, and an anomaly detector flags unusual or possibly misrouted cases for human review. Instead of just spitting out an answer, it shows the recommendation with a confidence score, the key factors behind it, and the rule-check result so a human (officer) always makes the final call. Down the line it can also connect people to verified legal-aid providers.
Innovation
JuaHaki is unique because it's not a chatbot or a fully automated system it's a three-layer decision-support tool. ML classifies the case, a rules layer checks it against real Tanzanian jurisdiction thresholds, and anomaly detection flags anything unusual for human review. Every recommendation comes with a confidence score and reasoning, so it's explainable rather than a black box. Unlike Tanzania's current manual court referrals or generic legal-info tools, JuaHaki engages with the specifics of each case and is built around actual local court structures (Primary Court, Ward Tribunal, DLHT, CMA), with legal-aid matching .
Expected Impact
JuaHaki's impact is mainly about cutting down the confusion and delay caused by cases landing in the wrong forum citizens get faster, clearer guidance on where their legal matter belongs and whether they need legal assistance, while court and tribunal officers benefit from quicker case triage, fewer avoidable referrals, and reduced backlog. Because every recommendation is checked against real jurisdiction rules before it reaches a human, it also improves consistency in routing decisions across the justice system. More broadly, it narrows the access-to-justice gap for people unfamiliar with how courts and tribunals work, without replacing legal professionals it simply helps people and institutions get to the right starting point faster.