We design and build the AI and data platforms behind modern Saudi banking: fraud and risk, lending decisions, customer intelligence, Arabic assistants and open APIs, all connected to the systems you already run and built to stand up to SAMA scrutiny.
Platforms sit beside your core banking system and talk to it through APIs, events or batch, so modernisation does not wait on a multi-year core programme.
Inputs, scores, reasons and outcomes are logged, so risk, audit and regulators can see how an automated decision was reached.
Right-to-left interfaces, Arabic dashboards and assistants that handle Gulf phrasing, reviewed with native speakers.
Start with one and add the rest on the same data foundation. Each sits behind the same governance, so adding a capability does not mean adding a new compliance problem.
Real-time fraud scoring, AML transaction monitoring with SAR case management, credit and market risk dashboards, and SAMA regulatory reporting generated from governed data.
Application-to-decision journeys for personal, SME and auto finance: data collection, SIMAH and alternative-data scoring, policy rules, affordability checks and automated approvals with a human-review path.
A single customer view that feeds next-best-offer, churn and lifetime-value models, so product teams can target by need rather than by blanket campaign, with consent and PDPL controls built in.
Banking assistants for app, web and WhatsApp that handle balances, card controls, disputes and product questions in Arabic and English, with secure authentication and a clean hand-off to an agent.
API gateways, consent management and developer portals that let a bank expose accounts, payments and data to partners and fintechs securely, with throttling, monitoring and clear audit logs.
The operational side of PDPL and AI governance: data discovery, consent, rights workflows, retention, breach readiness and approved-use rules for AI tools, designed to be evidenced to your counsel and regulators.
Each release goes live and is measured against a baseline before the next begins. It keeps programmes honest and gives your board real results to look at.
Most failed AI projects fail on data quality and access, not algorithms. We prove the data pipeline first and treat it as a product in its own right.
SAMA, SAFIU and PDPL requirements shape the architecture from week one rather than being checked at the end. Your compliance team reviews as we build.
Runbooks, documentation and training are part of delivery, so your analysts and engineers can operate and extend the platform without us.
Platforms connect to core banking, card and payment switches, credit bureau data, CRM and your existing data warehouse. Exact connectors depend on your vendors, and we confirm the interface approach during discovery.
We map systems, data and pain points, rank use cases by value and readiness, and agree how success will be measured.
Data pipelines, integration and governance controls are put in place for the first use case.
One platform goes live alongside existing controls, then is tuned against the baseline with your risk and business teams.
Further channels and platform areas are added on the same foundation, with handover and training for your team.
Timelines depend on data readiness and integration complexity and are confirmed after discovery.
They are the software systems that let a bank use data and machine learning in daily operations: scoring transactions for fraud, deciding on credit applications, personalising offers, answering customers in Arabic and English, exposing services through APIs, and reporting to the regulator. Crux designs and builds these platforms and connects them to your core banking, payments and data systems.
No. We build around the core. Platforms read from and write back to core banking, card, payments and CRM systems through APIs, events or batch interfaces, so a bank can modernise decisioning, analytics and customer experience without a core replacement programme.
We design against the SAMA Cyber Security Framework, SAMA and SAFIU requirements for AML/CFT, Basel III and IFRS 9 reporting needs, and the Personal Data Protection Law. Your compliance and legal teams own interpretation and sign-off, and we provide the controls, logs and evidence they need.
They should be. Platforms record the inputs, score, contributing factors and final decision for each automated outcome, keep model documentation and validation records, and monitor drift. That lets a risk team show how a model behaves and why a specific decision was made.
Yes. Dashboards, customer journeys and conversational assistants are built for Arabic and English, including right-to-left layouts and Gulf Arabic phrasing for assistants, and reviewed with native speakers before release.
Start with the area where the cost is clearest and the data is most ready, often fraud losses, credit decision turnaround or manual regulatory reporting. We run a short discovery to rank candidates, agree how success will be measured, and deliver one release into production before expanding.
A short discovery call is enough to rank the options, name the data you would need, and sketch what a first release could look like.