The Rules of Prompting the AI Persona

Last updated: April 15, 2026

Writing rules for Conversation Context, Persona Background Info, and Persona Concerns

The AI persona doesn't follow a script — it responds based on the character you describe. Your word choices and sentence structure directly shape its behavior. These rules give you the foundation to write prompts the AI can interpret correctly.

Know the Two Identifiers

Your platform recognizes two reserved terms. Use them consistently across all prompt fields:

Term	Refers To	How to Use It
<persona name>	The AI character in the simulation	Always third person — describes who they are, how they think, what they feel
trainee	The human learner in the simulation	Named explicitly whenever the persona is reacting to, waiting on, or expecting something from the user

You can describe the trainee's job title — just anchor it with the word "trainee" first, then use "trainee" for the rest of the prompt.

Example: "The trainee is an Account Executive conducting a renewal call with <persona name>."

The Rules — By Prompt Section

Conversation Context — Keep the AI in the Dark

Describe the situation and why the meeting is happening. Do not reveal the trainee's goals or plan. If the AI knows the agenda in advance, it will anticipate moves and react before they happen.

✅ Situation only — AI must react to whatever the trainee brings
The trainee is a logistics account manager on a first call with <persona name>,
a procurement lead who responded to an outbound email. They have not spoken before.
❌ Gives the AI the roadmap — removes the challenge for the learner and tells the persona the trainee’s intentions
The trainee will ask about shipping volume and pitch a freight audit service.

Name the Object of the Sentence

When the persona is reacting to or waiting on the user, use the word "trainee" explicitly. Without it, the AI may assume that action is its own to perform.

❌ Avoid	✅ Use Instead
"<persona name> will only stay engaged if value is established."	"...if value is established by the trainee."
"Won't discuss budget without seeing a clear business case."	"<persona name> won't discuss budget unless a business case is presented by the trainee."

Persona Background Info — Write Narratively, Not Directively

This section is a character brief. Describe who the persona is and what has shaped their perspective — never what they will say or do. Directive language becomes an instruction; narrative language becomes a personality trait.

❌ Avoid	✅ Use Instead
"<persona name> will push back on anything vague."	"<persona name> has little patience for generic pitches and tends to probe for specifics before opening up."
"<persona name> needs data before agreeing to anything."	"<persona name> came up in an engineering background and questions claims that aren't backed by provable data."

Follow this pattern: "<persona name> tends to [behavior] because [underlying reason]."

Persona Concerns — Frame as Hesitations, Not Agendas

Concerns written as things the persona "will ask" or "needs to know" make the AI lead the conversation. Written as internal worries, they keep the AI reactive and the trainee in the driver's seat. Limit to 1–2 concerns.

❌ Avoid	✅ Use Instead
"<persona name> will ask about contract length and push for flexibility."	"<persona name> had a frustrating experience with a rigid contract at a previous company and is cautious about any long-term commitment."
"<persona name> needs to understand implementation before agreeing to anything."	"<persona name> is skeptical of solutions that sound simple in demos but end up requiring heavy lift from his team."

To control timing, add conditional language: "If pricing comes up before the trainee has established value, <persona name> will be noticeably more resistant." This prevents the AI from front-loading objections.

Quick Reference

Prompt Section	The Rule
Conversation Context	Keep the AI in the dark — describe the setup, not the trainee's plan.
Conversation Context	Name the object — use "trainee" when describing who must take action.
Persona Background Info	Write narratively — who the persona is, not what they will do.
Persona Background Info	Use "hesitant until" framing — set conditions the trainee must satisfy.
Persona Concerns	Frame as internal worries — hesitations the AI holds or challenges they deal with
Persona Concerns	Use conditional language — signal when a concern should surface, not just that it exists.

The goal is a conversation where the trainee drives and the AI responds — not the other way around. When in doubt, make it more narrative and less directive.