Configuration & validation
Item types, workflows, forms, dashboards, permissions: in Avera they're configuration, not code. No customization project, no admin army, and no system that freezes the day it's validated.
Items, attributes, relationships, lifecycles, layouts, permissions: the meta-model is configuration. A new regulated domain is a package, never a release.
Your item types and their fields, typed and validated, including calculated fields driven by formulas.
Which links may exist between which types, with direction and bounds; traceability is shaped, not free-form.
States, transitions, approval rules, endorsements, and signature requirements, per type.
Forms, worksheets, hubs, and dashboards, arranged per item type and per role.
Roles and grants down to individual fields in individual states; denied by default.
Every configuration change, from a new field to a whole item type, moves through one governed pipeline. Configuration has the same discipline as your design history.
01
A typed change: a new field, a reworked workflow, a whole item type. From an admin, or drafted by AI.
02
Changes collect in a change request. Nothing touches the live configuration.
03
A dry run against your real project data: what's affected, what would break, before anything applies.
04
A named human approves. Promotion is atomic and audited, like any controlled change.
05
The verification matrix is recorded for the new baseline; your validation statement shows drift the moment it exists.
Validation isn't a phase at the end. Every promotion leaves its evidence behind: that's what integrated CSV/CSA means.
Every promotion records the full verification matrix: every check, pass and fail, re-runnable on demand.
Issue a statement over a validated configuration: signed, with drift visible the moment later changes land.
Accepted findings are acknowledged with a rationale, not hidden: the honest state of validation, continuously.
A template package carries everything: a full QMS, a device template with design controls, risk, and V&V, test management, a validation kit. Configured to your program on arrival, imported through the same change control, and fully yours to evolve. Export everything the same way; no lock-in.
D0 · Operational in days
Foundations included, AI drafting the configuration to your process, and every change reviewed and promoted through change control: stand-up stops being a project. How AI configures the platform →
After go-live
The same mechanism keeps working after go-live, so continuous improvement is real, low-cost iteration, not theater.