Brainforge Knowledge

2026-04-10_eden-client-call-presentation-structure

4 min read

Eden client call presentation structure (EDEN AI only)

Date: 2026-04-10
Scope source: Linear EDE3 team board

Slide 1 - Reset and clear the air

  • Open by acknowledging the authentication back-and-forth directly.
  • Position this as friction that led to a better architecture than originally scoped.
  • Tone: calm, accountable, and forward-looking.

Suggested opener:

Before we jump in, I want to address the authentication back-and-forth head on. That was real friction, and we took it seriously. The upside is we implemented a better architecture than originally scoped: three separate instances - one for Danny, one for Adam, and one internal for our testing - with a clean promotion path across all three. That gives each of you your own data context and gives us a safer, faster release process.

Slide 2 - What changed in authentication setup

  • Previous setup: one shared instance tied to one account.
  • New setup: three instances
    • Danny instance
    • Adam instance
    • Internal Brainforge testing instance
  • Why this is better:
    • No single-account bottleneck
    • Better test reliability
    • Cleaner deployment and validation process

Slide 3 - Operational benefit of the new model

  • Each stakeholder can test using their own account and data context.
  • Brainforge can validate changes internally before promotion.
  • Updates can be synced across all three instances consistently.

Slide 4 - Live readiness checkpoint (decision gate)

  • Confirm Adam status: already set up and tested.
  • Ask Danny for progress based on Loom walkthroughs.
  • Branch in real time:
    • If Danny is ready → continue to progress update.
    • If Danny is blocked → pause and complete setup during the call.

Slide 5 - What we accomplished so far (EDEN AI only)

Completed highlights to show:

  • Built chat interface and deployed for M1 demo readiness.
  • Connected Mastra to Slack MCP and wired PII processing.
  • Built run_gws and gws_schema tooling with command whitelist and Docker setup.
  • Implemented identity resolution mapping and table seeding workflow.
  • Built PII redaction fixtures and validation suite.
  • Completed Mastra agent scaffold, tool wiring, and observability.
  • Completed production system prompt for GWS command routing and Slack handling.
  • Completed cross-surface validation of run_gws and command reference finalization.
  • Completed key access foundations:
    • GCP project/service account setup path
    • DWD approval process path
    • Slack app setup
    • Slack MCP feasibility spike
  • Defined identity mapping schema and shared test environment baseline.

Slide 6 - Core PII redaction review (internal review focus)

  • Core PII redaction function is still in internal review.
  • Current concern: redaction may be too sensitive and may be removing useful context.
  • Discussion point for client:
    • We are reassessing how much redaction is actually necessary given the integration/BAA setup.
  • Proposed product direction:
    • Keep strict redaction available, but evaluate making redaction level configurable in the future UI (for example, as a toggle/policy control).

Slide 7 - What gets us to org-wide querying

  • Current state: strong user-scoped foundation with security controls in place.
  • Next state: Google and Slack access patterns expanded toward organization scope.
  • Framing: this is not a pivot, it is the planned expansion layer after foundational stability.

Response for expected question (“I thought we could query all Eden data already”):

Great question - we built the secure foundation first so the system is stable and safe per user context. The next step we are actively working on is expanding Google and Slack connections from user-scoped to organization-scoped behavior. That requires backend scope and permission-path adjustments, which are now in progress.

Slide 8 - Product UI vision discussion

  • Demo a sample UI beyond built-in Mastra Studio.
  • Show example concepts:
    • Prompt library
    • GWS and Slack toggles
    • Cleaner operator workflow
  • Goal: gather feedback on what to keep, remove, and prioritize.

Suggested agenda (short version)

  1. Reset and authentication update
  2. Instance readiness check (Adam + Danny)
  3. Completed work (EDEN AI)
  4. PII redaction review and policy direction
  5. Org-scope expansion path
  6. Sample UI walkthrough and feedback

Graph View