ORGANISATIONAL SPACES · FRAMEVR

A place where the work can be seen

We build each organisation a shared space in FrameVR: open in a browser or a headset, with six rooms, one for each habit.

Most AI work happens in private windows. A shared space gives it somewhere to be: a wall you can walk up to, a room where the team meets the agents, a board that updates itself. Frame’s own Learning Central shows what a frame can be when people learn together in it.

FIG 4 · SIX ROOMS

Front DoorNAMEWelcome, the organisation’s manifesto and the Wall: one card per team, agent and autopilot, with owner and autonomy level.

The FloorPLACETeam zones. Each agent’s card sits in its team’s zone with a live status board.

Control RoomBOUNDThe autonomy ladder, approvals waiting, the incident log, the risk partner’s desk.

Meter RoomMETERMonthly cost and value boards by agent and team, with alerts.

ClassroomTEACHLearning paths by role, short videos, practice prompts and office hours.

AssemblySHAREShow and tell, the monthly census and the quarterly retire-or-renew, on a stage everyone can join.

PACKAGES · LAUNCHING 22 APRIL 2027

Ways to start

TEMPLATES

Shared Floor kit

Agent card, registry, autonomy ladder, meter sheet, ritual run-sheets, learning-path builder and Frame-ready boards.

SPACE STARTER

Founding Floor

A branded six-room space with your agents on the Wall, plus your first floor walk and show and tell, facilitated. Four weeks.

SPACE TEAM

Team Floor

Everything in Starter, plus live boards for registry, meter and incidents, and three months of facilitated rituals.

ENTERPRISE

Many floors

A floor per division, integrations, a governance pack mapped to NIST AI RMF and ISO/IEC 42001, and train-the-steward.

INTEGRATIONS

Live boards

Minimal feeds (agent registry, cost totals, learning progress) published to boards in your space. Metadata only.

PARTNERS

White label

Templates and space blueprints for consultancies and L&D providers to brand and resell.

Prices are being tested with early conversations and design partners, and will be published at the launch.

THE DEMO SPACE · FRAMEVR

Walk the Us and Them floor

MICROSOFT LEARN

Visit the area →

#usandthem-decision-1-models

Decision 1

A decision model, not a chatbot: give it a state (text or JSON) and fixed options, and it returns a typed decision with a probability for each option. No free-form text, no written rationale.

Autonomy: a control point for agents at A2-A4 - it can decide continue, stop, retry or hand off to a person (our mapping).

MICROSOFT LEARN

azure / foundry / foundry models / how to / use foundry models microsoft decision

Visit the area →

SYSTEM 1 · SYSTEM 2

Two speeds of working with agents

Borrowed from Daniel Kahneman’s Thinking, Fast and Slow, as a metaphor: fast, low-stakes work with agents, and slow, deliberate work where a person decides.

#usandthem-system-1

System 1 · Fast and Intuitive

Quick asks, drafts, summaries, first answers. Low stakes, easy to check, easy to undo.

MICROSOFT LEARN

copilot / cowork / cowork faq

Visit the area →

#usandthem-system-2

System 2 · Slow and Deliberate

Multi-step, high-stakes, hard to undo: money, customers, people decisions, regulated data.

MICROSOFT LEARN

azure / ai foundry / openai / how to / reasoning

Visit the area →

Decision 1 is Microsoft-Decision-1, a decision model in Microsoft Foundry: it returns a typed decision and a probability for each option rather than generated text, which makes it a natural fast “System 1” control point for agents. Microsoft’s guidance is to set confidence thresholds and keep meaningful human review for decisions about people. Board images.

GUARDRAILS

Safe by design

The space shows metadata (names, owners, levels, totals), never the sensitive data agents work on.
Live boards read from a minimal published feed; no credentials in the space.
Members by email; guests view-only unless invited.
Every board also exists as a web page or PDF, for accessibility.

TALK TO US ABOUT A SPACE