Demo brain. Simulated product for a fictional client — see onboarding-status.

Product Context

One-paragraph purpose: what Umbrella Corporation the product does, who its users are, and the domain knowledge an AI session needs before reasoning about feature work.

Company background

Founded in 1968 as an agrochemical business, Umbrella Corporation diversified into pharmaceuticals through the 1980s and now operates as a group of three divisions: Applied Biosciences (R&D labs, the subject of this engagement), Umbrella Pharmaceuticals, and Umbrella Consumer Health. Group headquarters are in Raccoon City; Applied Biosciences runs six research sites — Raccoon City, Barcelona, Osaka, Cape Town, Vancouver, and Warsaw — each historically operating with a good deal of local autonomy over its own tooling, which is exactly why UMBRA ended up as six independently-customised instances rather than one shared system. Umbrella is a public company (mid-cap, listed) with roughly 14,000 employees group-wide; Applied Biosciences is the smallest of the three divisions by headcount. (source: Google Drive, “Umbrella × Loomery — SOW v1.2”; Granola note “Kickoff”, 2026-06-15)

Each division runs its own operational systems independently, but Hive OS is being watched at group level as a possible template for the other two divisions’ own legacy plant-ops software — see engagement §Group interest. (source: Granola note “Group Steering Committee — Q3 Review”, 2026-08-14)

What it does

Hive OS is a lab-operations platform for Umbrella’s Applied Biosciences division. It replaces UMBRA, an on-prem VB.NET/Oracle system running since 2003 at each of Umbrella’s six research sites independently, with a single hosted platform covering three areas: facility access (badge issuance, zone-level permissions, override approvals), sample inventory (chain-of-custody from collection through disposal), and compliance reporting (audit-ready exports for regulators). Phase 1 covers facility access and sample inventory; compliance reporting is Phase 2. (source: Google Drive, “Umbrella × Loomery — SOW v1.2”; Granola note “Phase 1 Scope Review”, 2026-07-30)

Users

• Lab technicians — log sample collection/transfer/disposal events throughout the day; need this to be fast enough not to interrupt bench work.
• Facility security staff — issue and review badge access, action override requests when a technician needs zone access outside their normal permissions.
• Compliance & Quality (Priya Anand’s team) — pull audit-ready chain-of-custody and access reports ahead of inspections; today this means manually reconciling paper override logs against UMBRA exports, which is the exact gap Hive OS closes. (source: Granola notes, “Discovery: Facility Access Workflows” 2026-07-02 and “Discovery: Sample Inventory & Compliance Reporting” 2026-07-16)

Domain notes

See glossary for quick definitions of the terms below.

• Chain-of-custody must be reconstructable from the system alone with zero paper backstops — this is the hard requirement behind most Phase 1 design decisions (see decisions).
• Manual overrides (a technician granted temporary access outside their normal zone permissions) are today logged on paper at each site and are the single weakest point in Umbrella’s current audit trail. Hive OS makes every override a first-class, timestamped record.
• Site independence: each of the six research sites currently runs its own UMBRA instance with site-specific customisations, which is why the migration is strangler-fig (site-by-site) rather than a single global cutover — see decisions.
• Raccoon City pilot site is the first cutover target, chosen because it’s the smallest site by headcount and shares a facilities team with the compliance office, making rollback verification easier. (source: Granola note “Phase 1 Scope Review”, 2026-07-30)