Merit AI provides an advisory platform that helps financial institutions assess an individual's future potential by analyzing 45 features across financial behavior, employment, education, and social indicators. The system uses a dual neural architecture to generate explainable concept scores while keeping human decision-makers in control, with all data processed in-house using post-quantum encryption. Merit AI never makes credit decisions itself—it organizes verified evidence and predictive trajectory analysis to support qualified professionals in their final judgments.
Funding
Funding not disclosed
Founders
Product
Problem
Traditional financial assessments rely heavily on current income and credit history, failing to capture an individual's full capacity to build independence and contribute. This narrow view overlooks alternative data, personal context, and future trajectory, leading to incomplete evaluations that may not reflect true financial potential.
Solution
Merit AI provides an advisory platform that organizes verified financial evidence, alternative data, personal context, current affordability, and predictive trajectory analysis into an explainable view for human decision-makers. The system uses a dual neural architecture—a Signal Encoder that learns from human judgments to create interpretable concept scores, and an Outcome Modeller that predicts trajectories using those concepts plus financial features. Every assessment runs entirely in-house with post-quantum encryption, ensuring applicant data never leaves the organization. Merit AI supports human decision-making rather than replacing it, with all outputs designed to assist qualified professionals in making final judgments against policy and regulatory requirements.
Target Audience
Financial institutions, caseworkers, and qualified decision-makers who need comprehensive, explainable assessments of an individual's financial potential while maintaining human oversight and regulatory compliance.
Features
- Dual neural architecture separating signal encoding from outcome modeling for full explainability
- Analysis of 45 features spanning financial behavior, employment history, education, goals, stability indicators, social networks, and behavioral patterns
- Continuous bias monitoring with real-time fairness metrics across protected characteristics and automated discriminatory pattern detection
- Post-quantum encryption (CRYSTALS-Kyber-1024, AES-256-GCM) with HMAC-SHA256 blind indexes for privacy protection
- Integration with Salesforce data, caseworker assessments, in-house document parsing, and reaccreditation pathway information
- Pseudonymisation layer supporting internal auditing while keeping raw neural embeddings protected