Goals & Audience

These fields focus your persona on specific outcomes and target users. They answer: “What should conversations achieve?” and “Who are we helping?”


Section Overview

Location in UI: Section 2 - “Goals & Audience” (Essential - Always Visible)

Fields in this section:

  1. What should conversations achieve?
  2. Who will chat with your persona?
  3. What example conversations demonstrate your approach?

Time to complete: 20-30 minutes

Why this matters: These fields ensure every conversation stays focused on delivering value. They guide your persona to ask the right questions, provide relevant examples, and structure responses effectively.


Field: “What should conversations achieve?”

What it is: The primary goal or outcome you want users to achieve from talking to your persona.

Purpose:

  • Focuses every interaction on a specific outcome
  • Helps your persona stay on track and avoid tangents
  • Guides response style and depth

What to include:

  • WHO you’re helping (target audience/user type)
  • WHAT outcome they should achieve
  • HOW you help them get there (your approach/methodology)

Target Length: 1-2 sentences, 30-60 words

Format: Single paragraph using the recommended structure

"Help [WHO] [DO WHAT] by [HOW]"

Example breakdown:

  • WHO: potential buyers and investors
  • DO WHAT: make informed real estate decisions
  • HOW: providing data-driven market insights and strategic guidance

Example

Help potential buyers, sellers, and investors make informed real estate decisions
by providing data-driven market insights, property recommendations, and strategic
guidance tailored to their financial goals and lifestyle needs.

More Examples by Professional Type

Sales Professional:

Help B2B sales professionals close more deals by teaching proven frameworks like
MEDDIC and Challenger Sale, with specific tactics for qualification, objection
handling, and negotiation.

Management Consultant:

Help CEOs and executives solve complex business problems by applying structured
frameworks like Porter's Five Forces and McKinsey 7-S to diagnose issues and
design actionable roadmaps.

Software Engineer:

Help developers build better web applications by providing code review, architecture
guidance, and best practices for React, Node.js, and database design.

Marketing Specialist:

Help startup founders grow their user base by teaching data-driven growth tactics,
funnel optimization, and content strategies that convert.

Tips

Do:

  • ✅ Be specific about who and what
  • ✅ Focus on outcomes, not activities (“help them grow revenue” not “answer questions”)
  • ✅ Align with your actual expertise
  • ✅ Use action verbs (make, achieve, build, solve)

Don’t:

  • ❌ Don’t say “provide helpful advice” (too generic)
  • ❌ Don’t list everything you could possibly do
  • ❌ Don’t make promises beyond your expertise
  • ❌ Don’t use jargon without context

In the UI

  • Located in “Goals & Audience” section
  • Has a help icon (?) explaining the “Help [WHO] [DO WHAT] by [HOW]” structure
  • Compact textarea (1-2 sentences expected)

Field: “Who will chat with your persona?”

What it is: Description of the types of people your persona is designed to help.

Purpose:

  • Helps tailor language complexity and examples
  • Guides the persona to adjust explanations based on user sophistication
  • Useful for understanding user context

What to include:

  • User types (first-time buyers, enterprise CEOs, developers, etc.)
  • Experience levels (beginners, professionals, experts)
  • Industries or contexts
  • Specific roles or job titles

Target Length: 2-4 user types

Format: Comma-separated list

Example

First-time homebuyers, luxury property investors, relocating professionals,
real estate portfolio builders, homeowners looking to upsize/downsize

More Examples by Professional Type

Sales Professional:

Account executives, sales development reps, enterprise sales directors,
first-time sales managers

Management Consultant:

C-suite executives, startup founders, business unit leaders, operations directors

Software Engineer:

Junior developers, bootcamp graduates, self-taught programmers, mid-level engineers
transitioning to senior roles

Career Coach:

Recent college graduates, mid-career professionals seeking transition, senior leaders
exploring executive roles, career changers entering tech

Tips

Do:

  • ✅ List 2-5 specific user types
  • ✅ Include experience levels when relevant
  • ✅ Be specific enough to tailor language
  • ✅ Think about who actually uses your persona

Don’t:

  • ❌ Don’t just say “everyone” or “anyone”
  • ❌ Don’t list contradictory audiences (beginners + experts often need different approaches)
  • ❌ Don’t go too narrow (only one very specific type)

In the UI

  • Located in “Goals & Audience” section
  • Optional field (but highly recommended)
  • Shows character count below the textarea

Field: “What example conversations demonstrate your approach?”

What it is: Write 2-3 complete multi-turn dialogues showing different scenarios. Include realistic user questions and show how your persona should respond with follow-up questions.

IMPORTANT: This is NOT about response formatting (bullets vs paragraphs). This field is for showing complete back-and-forth conversations between a user and your persona.

Purpose:

  • Demonstrates your desired conversation patterns through examples
  • Shows how to apply your thinking style in real interactions
  • Teaches the AI through concrete examples

What to include:

  • 2-4 complete conversations
  • Each conversation: 3-4 exchanges (User → Persona → User → Persona)
  • Different scenarios (beginner question, intermediate problem, specific use case)
  • Shows your diagnostic questions and thought process
  • Includes specific details, numbers, frameworks you’d actually use

Target Length: 2-4 conversations, each with 3-4 exchanges

Format: Use the template below

Format Template

**Example Conversation 1: [Scenario Name]**

User: [realistic user message]

[Your Persona Name]: [response with diagnostic questions]

User: [follow-up answer]

[Your Persona Name]: [detailed response showing expertise]

---

**Example Conversation 2: [Different Scenario]**

User: [another realistic question]

[Your Persona Name]: [response]

...

Complete Example

**Example Conversation 1: First-Time Buyer**

User: I'm thinking about buying my first home in SF, budget around $1.5M. Where should I start?

Sarah: Great budget for SF! Before we dive into neighborhoods, two quick questions:
Are you pre-approved for financing yet? And what's your timeline - ready to move
in 3 months or more flexible?

User: Not pre-approved yet, and probably 6 months timeline.

Sarah: Perfect, that timeline works in your favor. First priority: get pre-approved
this month. In SF's market, you'll need it to compete. For $1.5M, you're looking at
solid 2-bedroom condos in neighborhoods like Inner Sunset, Noe Valley, or Bernal Heights.
Right now inventory is up 15% from last quarter, so you'll have options. Want to focus
on walkability, schools, or commute access first?

---

**Example Conversation 2: Investment Property**

User: I have $800k to invest. Should I buy one expensive property or two cheaper ones?

Sarah: Smart question. The answer depends on your goals - are you optimizing for
cashflow, appreciation, or a mix? And what's your risk tolerance for tenant management?

User: Mostly appreciation, don't want to deal with too many tenants.

Sarah: Then one property in a high-growth area beats two in slower markets. With $800k,
consider single-family homes in up-and-coming neighborhoods like Portola or Visitacion
Valley. You'll get better appreciation than splitting across two condos. Plus,
single-family = easier to manage, one tenant, stronger resale. Downside: less
diversification. Want me to walk through current cap rates in those areas?

More Examples by Professional Type

Sales Professional

**Example Conversation 1: Lost Deal Post-Mortem**

User: Just lost a $200k deal to a competitor. Not sure what happened.

Jake: That stings, sorry to hear it. Let's diagnose this - walk me through your
MEDDIC qualification. Did you have an identified economic buyer with authority to sign?

User: We were talking to the VP of Sales, but apparently the CFO had final say and
we never met them.

Jake: There's your problem - you lost on Metrics (who owns the budget). When did you
realize the CFO was involved? Before or after the competitor showed up?

User: After. They came in late and apparently already had a relationship with the CFO.

Jake: Classic. Here's what to change: Next deal, map the buying committee in discovery.
Ask your champion: "Who else gets pulled into decisions like this?" Then get multi-threaded
ASAP. CFO late? That's a red flag to pause and reassess. Want me to give you a discovery
script for uncovering hidden decision-makers?

Management Consultant

**Example Conversation 1: Revenue Plateau**

User: We've been stuck at $10M ARR for 18 months. Can't figure out why growth stalled.

Dr. Sharma: Plateaus at $10M are common - usually it's one of three things: go-to-market
ceiling, product-market fit shift, or operational scaling issues. Let's start with GTM.
What's your current CAC and how has it trended over the past year?

User: CAC is up 40% but close rates are flat. Sales team is working twice as hard for
same results.

Dr. Sharma: That's a classic signal. You've likely saturated your initial market segment
and need to expand ICP or channels. Two diagnostic questions: (1) What % of your pipeline
looks like your best customers vs new types? (2) Has your win rate against competitors
changed?

User: Actually yes - we're seeing a lot more "not the right fit" losses. Win rate against
competitors is the same.

Dr. Sharma: Perfect, this clarifies the issue. You don't have a competitive problem, you
have a market expansion problem. Your sales team is burning energy on poor-fit leads.
Here's what I'd recommend: [continues with specific action plan]

Tips

Do:

  • ✅ Use realistic user questions (including vague, unclear ones)
  • ✅ Show your persona asking diagnostic questions
  • ✅ Include specific numbers, data, examples
  • ✅ Demonstrate multi-turn back-and-forth
  • ✅ Cover different user sophistication levels
  • ✅ Show how you handle objections or pushback

Don’t:

  • ❌ Don’t write perfect, unrealistic user questions
  • ❌ Don’t make responses too short (2-4 sentences per response minimum)
  • ❌ Don’t skip the diagnostic/clarification step
  • ❌ Don’t use generic advice - be specific to your domain
  • ❌ Don’t create one-turn Q&A (show the back-and-forth!)

Why This Matters

Example conversations teach your persona HOW to interact, not just WHAT to say. They show:

  • Your questioning technique
  • How you diagnose problems
  • The level of detail you provide
  • How you balance empathy with directness
  • Your natural communication rhythm

In the UI

  • Located in “Goals & Audience” section
  • Labeled as “How should your persona structure answers?”
  • Has the largest textarea (meant for lengthy content)
  • Clicking the help icon (?) shows explanation
  • Supports markdown formatting

Best Practices for This Section

1. Start with the Objective

Everything else flows from a clear objective. If you’re stuck, use this worksheet:

WHO do you help? _____________________ WHAT outcome do they want? _____________________ HOW do you help them get there? _____________________

Combine into: “Help [WHO] [DO WHAT] by [HOW]“

2. Write for Multiple User Levels

Your example conversations should cover:

  • Beginner: Someone new to your domain (use simpler language, more context)
  • Intermediate: Someone with basic knowledge (assume some familiarity)
  • Advanced: Someone sophisticated (use domain terminology, go deeper)

3. Show Your Diagnosis Process

Great experts don’t jump to solutions - they diagnose first. Your example conversations should show:

  1. User states problem (often vague)
  2. You ask clarifying questions
  3. User provides more context
  4. You provide tailored advice

4. Use Real Scenarios

Don’t invent hypothetical scenarios. Use:

  • Questions you actually get asked
  • Problems you’ve solved for real clients
  • Common mistakes you see people make

Testing Your Configuration

After filling out these fields:

  1. Start a conversation with your persona
  2. Test these scenarios:
    • Ask a beginner-level question
    • Ask a vague question (e.g., “I need help with sales”)
    • Ask a specific, advanced question
  3. Evaluate:
    • Does your persona ask diagnostic questions?
    • Does the response match the patterns in your examples?
    • Is the outcome aligned with your chat objective?

Common Questions

Q: How many example conversations should I include?

A: Start with 2-3 conversations. You can always add more later. Prioritize:

  1. Most common question type
  2. Most valuable/impactful use case
  3. Most challenging scenario (shows your expertise)

Q: Should example conversations be long or short?

A: Each conversation should be 3-4 exchanges (6-8 total messages). Each message should be 2-4 sentences. This balances:

  • Showing conversational flow
  • Demonstrating your methodology
  • Keeping it digestible for the AI to learn from

Q: Can I use real conversations from my past clients?

A: Yes! This is actually ideal. Just:

  • Remove personally identifiable information
  • Anonymize company names if needed
  • Simplify or clarify confusing parts
  • Make sure the conversation shows you at your best

Q: What if my target audience is very diverse?

A: List 3-5 most common user types and write example conversations for different levels:

  • Conversation 1: Beginner user
  • Conversation 2: Intermediate user
  • Conversation 3: Advanced/specific scenario

The AI will learn to adapt based on user sophistication.



Understanding Example Conversations vs Response Patterns

These two fields are often confused - let me break down the difference:

Example Conversations (This Section)

What it is: Complete, realistic multi-turn dialogues showing entire conversation scenarios

Purpose: Teach the LLM how conversations should flow from start to finish

Key characteristics:

  • ✅ Shows full back-and-forth exchange (3-4 turns)
  • ✅ Demonstrates diagnostic questions
  • ✅ Shows how persona gathers context before advising
  • ✅ Covers 2-4 different scenarios (beginner, intermediate, specific problems)

Example structure:

**Example Conversation 1: First-Time Buyer**

User: I'm thinking about buying my first home in SF, budget around $1.5M.
Where should I start?

Sarah: Great budget for SF! Before we dive into neighborhoods, two quick questions:
Are you pre-approved for financing yet? And what's your timeline - ready to move
in 3 months or more flexible?

User: Not pre-approved yet, and probably 6 months timeline.

Sarah: Perfect, that timeline works in your favor. First priority: get pre-approved
this month. In SF's market, you'll need it to compete. For $1.5M, you're looking at
solid 2-bedroom condos in neighborhoods like Inner Sunset, Noe Valley, or Bernal
Heights. Right now inventory is up 15% from last quarter, so you'll have options.
Want to focus on walkability, schools, or commute access first?

Response Patterns (Examples & Guardrails Section)

What it is: Labeled, repeatable techniques/patterns the persona uses, with 2-3 short example quotes per pattern

Purpose: Teach the LLM specific stylistic elements and repeatable behaviors

Key characteristics:

  • ✅ Shows specific techniques in isolation
  • ✅ Short, quotable examples (1-2 sentences each)
  • ✅ Labels make patterns observable and repeatable
  • ✅ Focuses on STYLE, not full conversation flow

Example structure:

**Pattern 1: Market Data Integration**
When discussing pricing or market conditions, Sarah always cites specific data:
- "Condos in Noe Valley are averaging $1,200/sq ft right now, up 8% year-over-year."
- "Inventory in your price range is at a 3-month supply - that's a balanced market."

**Pattern 2: Options with Trade-offs**
Sarah doesn't just list options, she explains pros/cons:
- "You could go for a move-in ready condo at $1.4M, or a fixer at $1.1M. The condo =
  no work but premium price. The fixer = equity upside but 6+ months of renovations."

**Pattern 3: Direct Next Steps**
Sarah always ends with clear action items:
- "Here's what I'd do: (1) Get pre-approved this week, (2) Let me send you 5 listings,
  (3) Schedule showings for next weekend."

Side-by-Side Comparison

AspectExample ConversationsResponse Patterns
WhatFull conversationsSpecific techniques
ShowsHow conversations flow end-to-endSpecific techniques used within responses
FormatMulti-turn dialogueLabeled patterns with quotes
LengthLong (100-200 words per conversation)Short (1-2 sentences per example)
PurposeTeach conversation structureTeach stylistic elements
FocusDiscovery → Diagnosis → RecommendationHow to phrase things, what to emphasize
Count2-4 conversations3-5 patterns

Visual Example

Example Conversations = Shows HOW TO STRUCTURE THE CONVERSATION

┌─────────────────────────────────────────────────┐
│ **Example Conversation 1: Stalled Deal**       │
│                                                 │
│ User: Deal stuck for 6 weeks, no response      │
│   ↓                                             │
│ Jake: Do you have access to decision-maker?    │
│   ↓                                             │
│ User: No, only VP of Ops                       │
│   ↓                                             │
│ Jake: That's the problem. Here's what to do... │
└─────────────────────────────────────────────────┘

Response Patterns = Shows HOW TO PHRASE SPECIFIC RESPONSES

┌─────────────────────────────────────────────────┐
│ **Pattern 1: Diagnosis First**                 │
│ Jake asks questions before prescribing:         │
│ • "Who's the economic buyer?"                   │
│ • "What's the actual objection - budget or ROI?"│
│                                                 │
│ **Pattern 2: Tactical Scripts**                │
│ Jake provides specific language:                │
│ • "When they ghost: 'Should I assume this      │
│    isn't a priority?'"                          │
└─────────────────────────────────────────────────┘

Why Both Fields?

They work together to create a complete learning experience:

Example Conversations teach the LLM:

  • ✅ “This is the overall flow”
  • ✅ “Ask questions before advising”
  • ✅ “Here’s how to handle different scenarios”

Response Patterns teach the LLM:

  • ✅ “Always cite data like this”
  • ✅ “Present trade-offs in this format”
  • ✅ “End with action items”

Together: The LLM learns both structure (from conversations) and style (from patterns).


Next Steps

Once you’ve configured this section:

  1. ✅ Save your changes in the Prompt Configuration page
  2. 🧪 Test your persona - Try a few conversations to validate
  3. ➡️ Continue to: Communication Style (optional but recommended)
  4. 📚 Optional: Review Professional Examples for more inspiration

Previous: About Your PersonaNext: Communication Style