Persona Agent
An AI agent assigned a specific buyer or expert persona that evaluates content from that perspective during multi-agent review.
Persona agents are specialized instances of Claude loaded with a specific viewpoint: the skeptical VP of Sales, the technical evaluator, the budget-conscious CFO, the implementation-worried ops manager. Each agent has a persona file defining their priorities, objections, and evaluation criteria. In the STEEPWORKS content pipeline, 7 persona agents independently score articles before a moderated debate round. The key design constraint: persona agents must run via TeamCreate with persistent identity, not Task tool calls, because personas lose their memory between Task invocations. This was re-discovered twice before being codified.
Where it shows up:
Related Terms
Methodology
Multi-Agent Debate
A moderated discussion between persona agents where each argues from their buyer perspective, produc...
Agent Layer
Agent
A skill with autonomy. Agents run on schedules or triggers, execute multi-step workflows, and produc...
Methodology
VNC Framework
Visceral/Novel/Credible scoring framework that evaluates whether buyer-facing content triggers emoti...
Methodology
Quality Gate
An automated checkpoint in a skill chain that evaluates output against defined criteria before allow...