Start with a product, not a portfolio
Base Workers began by building Marasim: a Morocco-first wedding marketplace, professional operations platform and collaborative planning ecosystem. The product is in active implementation and pre-market validation.
That starting point matters. An application connects interface choices to real operational questions: what information is public, who can access private records and which external services can receive data. The requirements emerge from building the product.
Extract the privacy boundary
Public discovery and private operations need different boundaries. Customer, event, financial and contract information should not enter an anonymous marketplace search response simply because the same application stores it.
Connecting an application to cloud, SaaS, API and AI services introduces another boundary. PrivacyGuard grew from the need to inspect sensitive information, apply policy and transform or deny it before forwarding to an approved external destination.
PrivacyGuard is being hardened toward pilot preparation. Local synthetic implementation evidence is not the same as a production deployment or a guarantee of legal compliance.
Research the next question
Patterns and explicit field rules are a useful starting point for detection. Contextual classification raises a different question: can a system understand sensitive information without sending that information to a general-purpose external AI service?
Titrit is our research direction for structured decision intelligence. Contextual PII classification is its first strategic specialization for PrivacyGuard. A future classifier can inform a privacy system, but it should not get authority to change the policy. Titrit detects and classifies; PrivacyGuard decides and enforces.
Keep the products independent
These connections explain our strategy. They do not make the products a compulsory bundle. Marasim has a vertical application scope, PrivacyGuard has an infrastructure scope and Titrit has a research scope.
Our approach is Build → Learn → Extract → Productize. We build applications, discover recurring problems, develop the technology internally and test whether that technology deserves an independent product. We measure before scaling.