Agents
Agents
Brand Onboarding, Brand Crafter, BrandIQ, Campaign Crafter, Persona Builder. They share one voice.
NeoIQ has five named agents. They share one voice. They are named functions, not personalities.
Media placeholder. Animated GIF, ~25 seconds. Path:
/docs/media/agents-overview.gif. Shows: the four agent cards on the home
screen, a brief tour of each agent's primary action.
The five agents
| Agent | Role | Lives in |
|---|---|---|
| Brand Onboarding | Guides first-pass setup. Fills the Brand Kit shape from evidence. | Brand onboarding flow. |
| Brand Crafter | Maintains identity systems. Refreshes the canon. | Brand Brain refresh. Re-validation. |
| BrandIQ | Answers questions about the masterbrand. Search and retrieval. | Across every product area. ⌘K. |
| Campaign Crafter | Generates campaigns: concepts, briefs, content pieces. | Strategy Studio. Content Engine. |
| Persona Builder | Generates audience personas from data. | Strategy Studio. Open Canvas. |
What every agent does
- Drafts. Never produces without reversibility.
- Cites. Every claim names a source.
- Hands off. Output passes back to the operator. The agent never approves its own work.
- Refuses. If a request falls outside the canon's permissions, the agent says no with a named reason.
What every agent does not do
- Approve. Approvals name a human, never the model.
- Speak in the first-person plural. The agent says "I drafted three options," not "we drafted."
- Bleed across brands. An agent operating in Brand A's context does not surface Brand B's canon.
- Write to Brand Brain without the operator's named approval.
The two voices
Every agent operates in two voices:
- Itself, inside NeoIQ. Calm, named, terse. Cites sources. Passes the baton back.
- The client's voice, when producing client work. A campaign for the brand reads as the brand. The agent is invisible in the deliverable.
This is the entire product. The agent voice never bleeds into the client deliverable.