Trust, by design.
Rolling out a platform that states how your organisation works usually means that you only get the exec POV. Metacosm is built so that you only ask once, but you ask everyone: people, systems, technology, and workflows.
How work gets done today becomes plain to see, and the observation phase that keeps it current arrives with its rules set before the first integration lands. No sensitive data slipping through the cracks in the name of "getting set up".
The studio holds what you author
Today's product is an authoring studio: structure, ownership, dependencies, skills, and experience, entered deliberately by you and your team. Nothing in it observes anyone working. There are no integrations reading your tools, no activity feeds, and no scores derived from behaviour — skills are recorded as experience, never computed from surveillance.
What does exist is accountability for the tool itself. Every edit is a command in a hash-chained, append-only log, so the graph is replayable and auditable from day one. Workspaces are isolated down to the database row. And when our own support staff act inside your workspace, a banner says so and every action is recorded.
What you can tell your people
Reading signals from your delivery tools is a later phase, and these rules are fixed before the first integration lands. Each is a property of how the system will collect and hold data, checkable from inside the product. When your workforce asks what this thing can see, these are the answers.
- 01
Who did what is never kept
Metacosm knows who sits where; the org chart is part of the map, and team-level readings depend on it. What it will never keep is who did what. Observed work loses its individual attribution as it is collected, and findings attach to teams, roles, and systems.
- 02
Small groups stay unreported
Below a minimum group size, Metacosm shows nothing at all. It will not estimate, infer, or fall back on "anonymised" data.
- 03
What people share stays collective
Workforce interviews and surveys strengthen or weaken the overall picture. Nothing anyone says is shown as their words.
- 04
No finding from a single signal
One data point is not a finding. Nothing surfaces until independent evidence agrees.
- 05
Findings point at structure
Every finding is about how the organisation is put together, with fixes to match: redesign, cross-train, document, rebalance.
- 06
Teams see findings about themselves first
Nothing about a team goes up the chain before the team itself has seen it.
- 07
No customer content retained
The platform looks at how work moves through your organisation. Customer conversations and content are never stored.
- 08
The limits are inspectable
Anyone in the organisation can see what's connected, what can be seen, and where every limit sits.
- 09
Every finding can be challenged
Challenge a finding from inside the product, and the challenge becomes part of the record.
Absent — not denied
The first rule has a consequence worth spelling out. Ask what the system has observed about a person — how fast they work, what they wrote, how they compare — and the answer comes back absent, not denied. The data is discarded as it is collected, so there is nothing in the system to return.
The same rule applies to machines. An AI agent issuing ten thousand queries gets no further, because absent data cannot be reconstructed by asking again.
A shorter rollout conversation
Bringing in a tool that touches how work happens normally means a long conversation with your People team and your workforce. In Australia it also means workplace-surveillance notice laws, the Privacy Act, and WHS duties that treat intrusive monitoring as a psychosocial hazard. Metacosm keeps that conversation short: today the product holds only what you put in, and observation arrives with its rules already set.