Skip to main contentSkip to footer
Closed Beta

Govern agent workflows before they reach production.

This closed beta gives invited teams access to the AI Design Blueprint validation workspace and authenticated MCP tools. The goal is not a generic product tour. The goal is to test whether shared doctrine, injected directly into local agent tooling and review workflows, improves how teams build, validate, and govern AI-assisted work.

Continue to beta sign-inFor agents setup guidance

Access is limited to allowlisted participants. Please sign in with the same email address that received this invitation.

What this beta is testing

AI Design Blueprint is designed to bring product, design, and engineering into a shared governance model for agentic systems. This beta is focused on three practical questions:

  • ·Can doctrine injected through MCP improve the quality of AI-assisted output?
  • ·Can teams review agent behaviour more clearly through explicit handoffs, approval boundaries, and visible workflow state?
  • ·Can internal company rules, design systems, workflow constraints, architectural standards, be layered into the validation process in a usable way?
Who this beta is for

Invited practitioners working on real AI-assisted delivery workflows, especially in teams where product, design, and engineering need a shared way to review and govern system behaviour.

  • Engineers working with agent tooling in real workflows
  • Product managers reviewing AI-assisted execution paths
  • Designers shaping trust, visibility, approvals, and interaction behaviour
  • Cross-functional teams testing AI output against internal standards
Week 1 mission

The first week is intentionally bounded.

  1. 1

    Sign in to the beta environment

    Use your allowlisted email address. This gives you access to the beta workspace and authenticated MCP tools.

  2. 2

    Connect the AI Design Blueprint MCP

    Add the MCP endpoint to your preferred local agent environment. Setup guidance is on the For Agents page.

  3. 3

    Run one real workflow

    Use the protected MCP in one real task, build a component, review an existing workflow, or validate an agent-driven interaction.

  4. 4

    Ask the agent to validate against the doctrine

    Explicitly prompt the agent to validate the architecture against the AI Design Blueprint doctrine. Observe whether it changes behaviour.

  5. 5

    Report friction directly

    Reply to the beta invitation email. This beta is being managed on a direct 1:1 basis.

Included in this beta phase
  • Access to the beta validation workspace
  • Authenticated MCP tools for invited participants
  • Guided testing of one real workflow
  • Optional custom context overlay (reply to beta email with your team's rules)
  • Direct 1:1 feedback channel during the beta period
Optional: test against your own internal standard

The public doctrine is only the baseline. If your team wants to test the workflow against internal standards, design system guidance, architectural rules, workflow constraints, or review criteria, reply to the beta email with those materials. Where appropriate, a custom context overlay can be prepared for your workspace.

Access conditions

This is a closed beta. Access is limited to invited participants and allowlisted email addresses. If you were invited, please use the same email address that received the beta invitation. If you were not invited, this page does not currently provide open access.

Continue to your beta environment

If you received an invitation, sign in with your allowlisted email address to begin Week 1.

Continue to beta sign-inFor agents setup guidance

AI Design Blueprint, early-stage platform for trustworthy agentic workflows and production AI design standards.