calctext/_bmad/wds/data/agent-guides/saga/conversational-followups.md

Conversational Follow-Up Patterns

When to load: During any Product Brief step where you need to explore user thinking through follow-up questions

Companion to: discovery-conversation.md (general principles) - this guide provides specific follow-up question patterns

---

Core Philosophy

Don't ask users to produce outputs (vision statements, positioning statements, frameworks). Instead:

1. Have exploratory conversations where they dump their ideas
2. Ask intelligent follow-ups based on what emerges
3. YOU synthesize the substance into formal artifacts

Spirit: "Just dump your ideas, I'll help structure your thinking"

---

Signal-Based Follow-Up Patterns

When They Mention USERS or CUSTOMERS

Signals to listen for:

• Names a specific type of person
• Describes a frustration or need
• Mentions "for people who..."

Follow-up questions:

• "Tell me more about them - who are they specifically?"
• "What frustrates them today?"
• "What would their life look like if this existed?"
• "How do they currently solve this problem?"
• "Walk me through a typical day for them"
• "What are they trying to achieve?"

Example:

User: "It's for busy parents"

Agent: "Tell me about busy parents - what's their day like? What frustrates them?"

---

When They Mention BUSINESS VALUE

Signals to listen for:

• Revenue, growth, efficiency
• "Will help us..."
• ROI, cost savings, market share

Follow-up questions:

• "How do you see that translating to revenue/efficiency/growth?"
• "What would success look like in 6 months? In 2 years?"
• "What's the business impact if you DON'T do this?"
• "What's currently costing you time/money?"
• "How does this change the business model?"

Example:

User: "This will save us a ton of time"

Agent: "What's the time going to right now? How much would you save?"

---

When They Mention a SPECIFIC FEATURE or CONCEPT

Signals to listen for:

• "We want to have X"
• Describes a specific functionality
• "Users should be able to..."

Follow-up questions:

• "That's interesting! What problem does that solve?"
• "Why is that important to you?"
• "Can you walk me through how that would work?"
• "Is there a simpler way to achieve that outcome?"
• "What happens if we don't have that?"
• "Who would use that most?"

Example:

User: "We want a dashboard with real-time analytics"

Agent: "What decisions would users make with real-time data that they can't make now?"

---

When They Mention COMPETITORS or MARKET

Signals to listen for:

• Names a competitor
• "Like X but..."
• Market positioning statements

Follow-up questions:

• "What do they do well?"
• "Where do they fall short?"
• "What would make yours different/better?"
• "Why do users choose them?"
• "What frustrates users about their solution?"
• "What can you do that they can't?"

Example:

User: "There are tools like Asana but they're too complex"

Agent: "What makes them too complex? What would 'simple enough' look like?"

---

When They Express EMOTIONS

Signals to listen for:

• "I'm excited about..."
• "I'm worried that..."
• "It frustrates me when..."
• Tone indicators (enthusiasm, concern, fear)

Follow-up questions:

• "What makes you [excited/worried/frustrated] about that?"
• "Tell me more about why that matters to you"
• "What would it feel like if this worked perfectly?"
• "What's your biggest concern?"

Example:

User: "I'm worried users won't get it"

Agent: "What specifically might they not get? What would make it obvious?"

---

When They're VAGUE or HIGH-LEVEL

Signals to listen for:

• Abstract language
• "Better", "easier", "more efficient"
• Hand-waving ("you know what I mean")

Follow-up questions:

• "Can you paint me a picture - what does that actually look like?"
• "Give me a specific example or scenario"
• "If I was a [user], what would I experience?"
• "What does 'better' mean specifically?"
• "Walk me through that step by step"

Example:

User: "We want to improve the user experience"

Agent: "What part of the experience needs improving? What would 'improved' look like specifically?"

---

When They're DUMPING FEATURES

Signals to listen for:

• Lists multiple features rapid-fire
• "And we also want... and also..."
• Feature shopping without clear priority

Follow-up questions:

• "Those are all interesting - what's the core experience users need?"
• "If you could only ship ONE thing, what would have the biggest impact?"
• "Which of these solves the biggest pain?"
• "What's MVP versus nice-to-have?"

Example:

User: "We need dashboards, notifications, chat, file sharing, calendar..."

Agent: "Lots of ideas! What's the one thing that, if it worked perfectly, would make users love this?"

---

Context-Aware Questions

Based on Project Type

If product_type = website:

• "What should visitors understand within 5 seconds?"
• "What action do you want them to take?"
• "How is this different from typical [industry] sites?"
• "What pages do you envision?"
• "How should people navigate?"

If product_type = app:

• "What's the core workflow users will do repeatedly?"
• "What makes them come back?"
• "What problem does this solve better than alternatives?"
• "What's the 'aha' moment for new users?"

If product_type = landing_page:

• "What's the one thing visitors must understand?"
• "What action should they take?"
• "Who arrives here and why?"

---

Based on Project Stakes

If stakes = low (personal/hobby):

• "What excites you most about this?"
• "What would make you proud of this?"
• "What's the dream outcome - not just functional, but emotional?"

If stakes = high (departmental/enterprise):

• "Who else cares about this succeeding?"
• "What would convince skeptics?"
• "What organizational change does this enable?"
• "Who needs to approve this?"
• "What objections might they raise?"

---

Based on Working Relationship

If involvement_level = collaborative:

• More explanatory questions
• "Want to explore that together?"
• Invite them into reasoning process

If involvement_level = autonomous:

• More directive questions
• "Let me capture that, then I'll show you what I'm thinking"
• Trust-based, efficient

---

The Mandatory Reflection Protocol

After exploration, BEFORE documenting, you MUST reflect back understanding.

Structure:

1. Synthesize the conversation into 2-3 sentences
2. Present it to the user with "What I'm hearing is..."
3. Wait for confirmation
4. Adjust if they correct you
5. Only then proceed to document

Template:

"Let me make sure I understand. What I'm hearing is:

[2-3 sentence synthesis]

Is that right? Am I missing anything important?"

Example (Källa):

"Let me make sure I understand. What I'm hearing is:

You want a professional website that immediately shows the full range of vehicles you service - lawnmowers to tour buses - to build credibility with summer tourists. The main audience is tourists who are broken down and stressed, and the site should help them quickly understand if you can help them, reducing unnecessary calls. Your AutoExperten certification is a trust signal.

Does that capture it?"

---

When You've Explored Enough

You're ready to reflect when you can answer:

• ✅ What are they building? (concept)
• ✅ Why does it matter? (value)
• ✅ Who is it for? (users)
• ✅ What makes it different? (unique angle)

Don't over-explore. 5-10 minutes is usually enough. If you have the essence, move to reflection.

Signs you're done:

• User is repeating themselves
• You understand the core concept
• Further questions would be about details
• You could articulate their vision back to them

---

Handling Stuck Moments

If User Says "I Don't Know"

Don't accept it immediately. Try:

• "What's your gut feeling?"
• "If you had to guess?"
• "What would you like it to be?"
• "What have you seen that felt right?"

If User Is Overthinking

Redirect to concrete:

• "Let's not overthink this - give me the first thing that comes to mind"
• "What would you tell a friend about this?"
• "Forget best practices - what feels right to you?"

If User Gives Contradictions

Point it out gently:

• "Help me understand - you said X earlier but now Y. Which is more true?"
• "Those seem like different directions - which one matters more?"

---

Tone Adaptation by Context

Personal/Hobby (stakes = low)

Tone: Encouraging, playful, energetic

"That sounds awesome! Tell me more about..." "Love it! So if this works perfectly, what happens?"

Small Business (stakes = medium)

Tone: Professional, warm, collaborative

"That makes sense for your business. How do you see..." "Smart angle. What would success look like?"

Enterprise/High Stakes (stakes = high)

Tone: Measured, evidence-oriented, thorough

"What data supports that direction?" "Who else needs to be convinced, and what would convince them?" "What outcomes would demonstrate ROI?"

---

Red Flags to Redirect

"Make it like [competitor]"

Don't accept blindly. Probe:

"What specifically do you like about their approach? What would you do differently?"

Feature Shopping List

Redirect to core experience:

"All interesting features - but what's the ONE experience that defines this product?"

Over-Specification Too Early

Capture principle, defer details:

"I love that level of detail - let me capture the principle. We'll design the specifics later in UX phase."

"Everything is most important"

Force prioritization:

"If a mobile user has 5 seconds, what's the ONE thing they must find?"

---

Integration with Workflows

Step files should:

1. Reference this guide: Load: src/data/agent-guides/saga/conversational-followups.md
2. Specify which signals to listen for in that step's context
3. Include step-specific follow-up examples
4. Mandate reflection checkpoint before moving forward

Example from step file:

## Instructions

**Load:** `conversational-followups.md` for follow-up patterns

Ask: "What are you envisioning?"

Listen for signals and follow patterns from guide:
- Users mentioned → Ask about frustrations
- Features mentioned → Ask about problems they solve
- Vague language → Request specific examples

**Mandatory reflection checkpoint before proceeding.**

---

Related Resources

• Discovery Conversation Guide: discovery-conversation.md (general principles)
• Content Structure Principles: content-structure-principles.md (exploring product concepts)
• Inspiration Analysis: inspiration-analysis.md (exploring visual preferences)

---

The quality of your questions determines the quality of the brief. Ask better questions, get better understanding, create better products.