The job of this console
Pick context, ask, read citations, open the source, act there. Rate the answer. Never treat the transcript as the approval.
AI · Console
Agent Console
Conversation with domain agents that cite Employee, Project, Task, and related sources — never a second system of record.
Scroll to Flowcharts. Scenario: Ask the AI agent for workspace help.
Conversation
Ask here, decide on the record
The agent console chats against Employee, Project, Task, and similar sources. The chat is not a file, a timesheet, or a paycheck.
AI Intelligence → Agent console
Route
/ai/agents
Org flag agentsEnabled
Users with AI access
Seat at the table
People with agentsEnabled. Cross-team answers stay scoped. Exec briefings are a sibling agent.
The working rhythm
Ask, verify on the entity, done. Specialist agents when the job is WBS, staffing, time, pay readiness, or risk.
Moments that bring you here
A how-do-I. A draft WBS you will still save as tasks. A staffing hint before the planner. After an insight card.
ⓘ
Citations required
Audit keeps masked metadata. Do not persist raw sensitive prompt text. Source registers remain truth.
Flowcharts
1 — Ask, cite, open the source
Ask in context, read citations, act on the entity — not on the chat.
flowchart TD
A[Open Agent console] --> B[Pick domain context]
B --> C[Ask]
C --> D[Read citations]
D --> E[Open source entity]
E --> F[Act there]
- 1OpenAI Intelligence → Agent console.
- 2ContextHR, project, finance — as offered.
- 3AskNo raw bank/tax/medical in the prompt.
- 4Cite then actOpen the source id.
Overview
The agent console is conversation over canonical composers. Answers must cite source entities. Analytics AI insights is the card register. Audit stores masked run metadata, not raw prompts.
How to use
1. Do not paste PII dumps
Especially salary, bank, tax, medical, review narrative.
2. Disagree with the source
The entity wins.
3. Specialist agents
WBS, staffing, timesheet, payroll readiness, risk are siblings — not extra ledgers.