Workspace / Support operationsCP
AI SUPPORT, WITH GUARDRAILS

Every reply.
Within reach.

A thoughtful support draft. A visible budget boundary. No surprises.

Live on Azure
◎
Real model, real budget

Every draft is a real Azure OpenAI call. A Microsoft Foundry Cost Control policy on the resource enforces a small per-minute USD budget and returns a real 429 when it runs out.

LIVE
MODEL CALLS
0

Real Azure requests this session

DRAFTS THIS SESSION
00

Only successful replies

OUTPUT GUARDRAIL
150 tokens

Maximum completion tokens

BUDGET SIGNAL
Waiting

From real response headers

THE 60-SECOND WALKTHROUGH

Meet your budget boundary.

1 Draft a real reply2 Watch the budget drain3 Hit the real guardrail
THE WORKBENCH

Turn a ticket into a next step.

HUMAN REVIEW REQUIRED
0 / 2,000

Memory only. Never saved or included in the frontend bundle.

Checking API configuration…

✧ Concise. Empathetic. Actionable.
SUGGESTED REPLYLIVE · REVIEW BEFORE USE
Your next great support reply starts here.
Enter the presenter code, then draft a reply or run the guided demo.
Always review before sending. Nothing is sent to a customer.
BOUNDARIES, NOT GUESSWORK

Budget radar

WAITING

Actual Foundry Cost Control headers · estimated, not billed cost

—available before call
— used— allotted
No call made yet

The radar fills from real x-ms-* budget headers returned by Foundry.

Model
Waking the scale-to-zero API…
Enforcement
Foundry Cost Control
Window reset
Not reported
Retry hint
—
Inspect budget evidence
No request yet.

The budget is enforced by Azure, not this app. It refills automatically each minute.

A CLEAR TRAIL

Activity stream

THIS SESSION ONLY
  1. Workspace ready

    Every draft is a real, budget-governed model call.