Evolution of a System to manage attendance of employees on Sites

Consolidate the scope, go-to-market strategy, multi-tenant architecture requirements, and data security standards for the execution of Claude-assisted development. 1. Executive Product Summary Presente (formerly operated internally as SmartTime) is a B2B SaaS solution focused on Proof of Presence, originally designed to resolve operational control of mobile personnel (commercial cleaning, maintenance services, security, and multi-location contract management). Core Value Proposition: Clock-in only works on site. The system guarantees, through facial biometrics and strict geofences in meters, that an employee can only register their attendance if they are physically within the authorized radius of the location. Key Differentiator: Real-time visibility of Failed Attempts (with exact distance in meters, interactive maps, and Mock GPS detection), eliminating fraud via impersonation (buddy punching or unauthorized remote clock-ins). 2. Market Strategy and Go-To-Market (GTM) Technical development must respond directly to commercial demands for the US market (initially focused on Florida and multi-site accounts): Ideal Customer Profile (ICP): Owners of commercial cleaning companies (janitorial/porter services), property maintenance, and security with 5 to 40 buildings or decentralized locations. Pain Point Solved: The administrator is not on-site; clients report absences, and contracts are lost due to lack of reliable evidence. Presente automates the "Location not opened" alert via WhatsApp and generates Visit Proof reports for third parties. Business Model: Tiered SaaS (Subscription per user/month or per site/month) with a short sales cycle based on 8 to 10-minute demonstrations focused on visual blocking of fraud. 3. Multi-Tenant Architecture (Core Requirements) To ensure scalability and strict commercial isolation, the architecture must comply with the following guidelines: Strict Data Isolation (org_id): All database-level queries (authentication, employees, check-ins, sites, and failed attempts) must be strictly partitioned and filtered by the organization identifier (org_id) to prevent any type of data leakage between tenants. Hierarchical Roles and Permissions: Owner / General administrator of the organization. Supervisor / Manager with visibility restricted to their assigned branches or zones. Employee / End user limited exclusively to clock-in actions on the mobile application. Dynamic Time Zone Management: Each site (branch/building) operates under its own time zone (e.g., Eastern Time, Central Time), which must be assigned automatically via geocoding when creating the location. 4. Data Security and Compliance (Data Privacy & Compliance) Given the handling of facial biometrics and geolocation, the system incorporates strict privacy policies: Biometric Consent: Mandatory flow for accepting terms and privacy policies at the moment of facial enrollment, registering the consent trace (biometric_consent_at). Sensitive Data Protection and Retention: Facial images from check-ins are stored in private buckets with signed URLs and configurable retention policies. Absolute prohibition of collecting government identification numbers (such as SSN) for US profiles. Device Integrity and Anti-Fraud: Android: Active detection of simulated locations (Mock Locations) and device integrity validation (Play Integrity / Root detection). iOS: Verification of consistency in horizontal precision, speed, and irregular movement heuristics (Impossible Travel). Modification Audit: Any manual adjustment to attendance records requires a justified exception (manual_adjustment) with an audit log of the responsible administrator. 5. Priority Technical Scope (Phase P0 - Demo-Ready) To guide development work (assisted by Claude in upcoming sprints), initial efforts will be concentrated on enabling critical demonstration capabilities: "Live" Operational Dashboard (Home): Replacement of static analytics with a live feed showing key real-time metrics (On site now, Late/missing, Failed check-ins, Sites covered). Meters and Geolocation Engine for Failed Attempts: Conversion of raw coordinates to exact distances in meters relative to the site radius, complemented by interactive comparison maps. Anti-Spoofing System: Automatic blocking of fraudulent check-ins with a red alert tag on the supervisor dashboard. Automatic Opening Alert: Operational cron job to evaluate grace windows by branch and trigger absence notifications via WhatsApp. Interface and Data Hygiene: Elimination of generic HR templates, standardization of phone prefixes (+1), and optimization of texts and language.