OPENAI BUILD WEEK · INTERACTIVE JUDGE SANDBOX

SalesView AI Operator

Operate the complete, human-controlled loop from sales context and GPT-5.6 judgment to test delivery, engagement signals, and the next reviewed action.

Sanitized judge demo

No external API calls or email delivery

WORKING PRODUCT WALKTHROUGH

Interactive operator cockpit

Synthetic data · Browser only
  1. 1 Context
  2. 2 Proposal
  3. 3 Approval
  4. 4 Delivery
  5. 5 Signals
  6. 6 Next action
Value hypothesis

Reduce manual registration follow-up for recurring community events.

Known constraint

Keep the first contact short and avoid promising a full system replacement.

Current next action

Review the personalized introduction before test delivery.

✦

Turn this context into a reviewable decision

This sandbox uses a precomputed, synthetic GPT-5.6 response. Production uses the OpenAI Responses API with Structured Outputs.

REAL PRODUCT BOUNDARY

AI judgment and deterministic delivery stay separate

01

SalesView

Reads sales context, asks GPT-5.6 for structured proposals, and keeps approval human-owned.

→
02

Mail Engagement

Owns Gmail delivery, idempotent retries, ownership checks, and engagement tracking.

↺
03

Next decision

Human opens, clicks, and replies return to SalesView for the next reviewed recommendation.

SAFETY AND PRIVACY

A functioning walkthrough without production access

  • Every name, message, delivery identifier, and engagement signal is synthetic.
  • No Google Sheets, OpenAI, Gmail, or Mail Engagement request is made from this page.
  • Controls update browser memory only and reset when the page reloads.
  • The authenticated production cockpit and all server API routes keep their access rules.