Company Name:
JURISTECH (JURIS TECHNOLOGIES SDN BHD) (618486-X)
Sector
Digital Services, Digital Finance, Telecommunication, Utility and Energy
Tech Enabler
Artificial Intelligence (AI) and/or Big Data Analytics (BDA)
Overview:
Introducing our automated Machine Learning (autoML) and AI platform leveraging advanced ML techniques to build powerful AI models. Tackling key issues in AI adoption, we streamline deployment, optimise algorithm selection, and enable continuous learning. Trusted by Malaysia's telcos and banks, it is used in credit management processes, providing benefits such as behavioural scoring, early non-performing loan detection (up to 6 months ahead), and identifying self-curing customers, optimising time and resources.
Our autoML and AI platform is used by banks and telcos in Malaysia to optimise and automate credit management processes. It can be used to:
- Digital Onboarding: personalise digital onboarding journeys
- Loan Application: provide behavioural scoring and alternative credit scoring for loan applications
- Debt Collection: predict non-performing loans and self-curing customers, identify behavioural late payers, provide personalised treatment journeys for delinquent customers, optimises collection efforts for financial institutions
- Fraud Detection: Offer fraud detection in loan applications through continuous learning
Additionally, our platform tackles AI adoption issues by:
- AI Deployment: simplifies AI deployment across industries
- Multiple Algorithms: provides optimal algorithm combinations for problem-solving
- Continuous Learning: ensures continuous learning for AI models
Pain Points:
- Digital Onboarding: Manual and time-consuming customer onboarding leads to long wait times and potential customer loss
- Loan Application: Limited credit access for unbanked/underbanked communities, making it difficult for banks to reach these segments
- Debt Collection: Manual and expensive debt collection processes prone to errors
- Fraud Detection: The industry shift towards digitalisation and automation has created opportunities for fraudsters to exploit vulnerabilities in digital systems
- AI Deployment: AI deployment is difficult and time-consuming
- Multiple Algorithms: One-size-fits-all approach is never optimised
- Continuous Learning: AI models that do not support continuous learning is costly and hard to upgrade
Solution/Services:
Juris Mindcraft
Our autoML and AI platform, covers everything from data pre-processing and machine learning modelling to deployment. It enables you to easily train, build, and deploy powerful models at scale, making intelligent predictions and business decisions effortless.
Our platform supports supervised learning (classification and regression), offers explainable models, and provides RESTful service deployment, enhancing coordination and speeding up deployment across data scientists, software engineers, IT teams, and business decision makers.
Key Features and Functionalities
- Loading Data: Drag and drop the dataset, select model metrics, and score the models.
- Data Pre-processing: Includes feature engineering, selection, transformation, data encoding, handling missing values, outliers, and checking variable types.
- Modeling: For modelling, you can select algorithms, tune hyperparameters, choose the top models, and perform model stacking.
- Evaluation: Use evaluation metrics, create plots and charts, compare models, analyse business value, conduct lifecycle analysis, and run champion-challenger tests.
- Deployment: Implement models in applications, convert to production-ready languages, test, through REST API.
- Monitoring: Maintain performance and reliability SLAs, detect performance declines, and rebuild models as conditions change.
Implementation:
Our autoML and AI platform seamlessly integrates with any existing infrastructure via API, requiring no customisation or configuration.
Benefit:
- Explainable AI - Explainable models and decisions that help you understand and interpret predictions.
- Easy Deployment - AI models deployed as RESTful service eliminating the discrepancy between different programming languages used by data scientists and software engineers.
- Uses the best of multiple algorithms - Creates a supermodel that is more robust, less prone to overfitting, and outperforms any single best-performing ML model.
- Continuous Learning - Ensures peak performance of production models despite changing conditions.
- Operationalises AIOps - Manages the full lifecycle of AI and decision models efficiently.