convoke-agents 2.0.0

package/_bmad/bme/_vortex/workflows/user-interview/steps/step-03-recruitment.md

@@ -0,0 +1,144 @@
+---
+step: 3
+workflow: user-interview
+title: Plan Participant Recruitment
+---
+
+# Step 3: Plan Participant Recruitment
+
+The most perfectly designed interview script is worthless if you talk to the wrong people. This step ensures you recruit participants who can actually give you the insights you need.
+
+## Why This Matters
+
+Recruitment is the most underestimated part of user research. Common mistakes:
+- Talking only to enthusiastic early adopters (survivorship bias)
+- Recruiting friends and colleagues who tell you what you want to hear
+- Getting participants who don't match your target user at all
+- Not recruiting enough people to see patterns
+- Spending weeks recruiting when you could start in days
+
+## Your Task
+
+Define your recruitment plan:
+
+### 1. Define Your Screening Criteria
+
+Who should you talk to? Be specific about must-have criteria and nice-to-have criteria.
+
+**Must-have criteria** (participant is disqualified without these):
+- What behavior or experience must they have?
+- What timeframe matters? (e.g., "used the product in the last 30 days")
+- What role, responsibility, or context must they be in?
+
+**Nice-to-have criteria** (for diversity, not disqualification):
+- Range of company sizes
+- Mix of experience levels
+- Different geographic regions
+- Different tools or workflows
+
+**Exclusion criteria** (who should you NOT talk to):
+- Competitors or people in your industry who would be biased
+- People with a financial relationship to your company
+- Friends, family, or close colleagues
+
+### 2. Determine Sample Size and Composition
+
+**How many participants do you need?**
+
+| Research Goal | Recommended Sample |
+|---------------|-------------------|
+| Exploratory discovery | 5-8 participants |
+| Hypothesis validation | 8-12 participants |
+| Comparative (two segments) | 5-6 per segment |
+| Usability testing | 3-5 participants |
+
+**Aim for diversity within your criteria.** If all 8 participants have the same background, you'll get one perspective eight times instead of eight perspectives once.
+
+### 3. Choose Recruitment Channels
+
+Where will you find these people?
+
+| Channel | Best For | Lead Time | Cost |
+|---------|----------|-----------|------|
+| Existing user base (email/in-app) | Current user research | 3-5 days | Free |
+| Social media / communities | Broad discovery research | 5-10 days | Free |
+| Professional networks (LinkedIn) | B2B research | 5-7 days | Free |
+| Recruitment panels (UserTesting, Respondent) | Fast turnaround, specific criteria | 1-3 days | $50-150/participant |
+| Customer support tickets | Users with specific problems | 2-4 days | Free |
+| Referral from other participants | Hard-to-reach segments | Ongoing | Free |
+
+### 4. Write Your Screening Survey
+
+Create 3-5 screening questions to qualify participants before scheduling.
+
+**Example screening survey:**
+> 1. What is your role? [Open text]
+> 2. How many people are on your team? [1-5 / 6-20 / 21-50 / 50+]
+> 3. Which project management tools do you currently use? [Checklist]
+> 4. When did you last evaluate a new project management tool? [Last month / Last 3 months / Last 6 months / More than 6 months ago / Never]
+> 5. Would you be available for a 30-45 minute video call in the next 2 weeks? [Yes / No]
+
+**Screening logic:** Disqualify anyone who answers [specific disqualifying answers]. Prioritize anyone who answers [ideal answers].
+
+### 5. Plan Incentives
+
+Incentives increase response rates and show respect for participants' time.
+
+| Participant Type | Suggested Incentive |
+|-----------------|---------------------|
+| General consumers | $25-50 gift card |
+| Professionals (B2B) | $75-150 gift card |
+| Senior executives / hard-to-reach | $150-300 gift card |
+| Existing happy customers | Sometimes free; product credit works |
+| Internal stakeholders | No incentive needed |
+
+**Alternatives to cash incentives:**
+- Donation to a charity of their choice
+- Early access to a new feature
+- Free subscription extension
+- Co-creation credit (name in release notes)
+
+### 6. Create Your Recruitment Timeline
+
+**Example timeline:**
+- Day 1-2: Post screening survey, send outreach messages
+- Day 3-5: Review responses, qualify participants, schedule interviews
+- Day 5-10: Conduct interviews (2-3 per day maximum to avoid fatigue)
+- Day 10-12: Follow-up interviews if needed
+
+**Important:** Never schedule more than 3 interviews in a single day. Interview fatigue degrades your listening quality after the third session.
+
+## Example
+
+**Screening Criteria:**
+- MUST: Small business owner (1-10 employees)
+- MUST: Currently uses spreadsheets for project tracking
+- MUST: Evaluated at least one PM tool in the last 6 months but did not adopt it
+- NICE-TO-HAVE: Mix of industries (not all tech)
+- NICE-TO-HAVE: Mix of team sizes within 1-10 range
+- EXCLUDE: Anyone who works for a PM tool company
+
+**Sample:** 8 participants
+
+**Channels:**
+- Primary: LinkedIn outreach to small business owner groups (free)
+- Secondary: Respondent.io panel with screening criteria ($100/participant)
+- Backup: Referrals from first 3 participants
+
+**Incentive:** $75 Amazon gift card
+
+**Timeline:** 10 days total (3 days recruiting, 7 days interviewing)
+
+---
+
+## Your Turn
+
+Please define your recruitment plan using the structure above.
+
+**Tip:** Start recruiting immediately after finalizing your plan. Recruitment always takes longer than you think. Send your screening survey today, not tomorrow.
+
+## Next Step
+
+When your recruitment plan is ready, I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/user-interview/steps/step-04-conduct.md

package/_bmad/bme/_vortex/workflows/user-interview/steps/step-04-conduct.md

@@ -0,0 +1,154 @@
+---
+step: 4
+workflow: user-interview
+title: Conduct Interviews
+---
+
+# Step 4: Conduct Interviews
+
+This step provides guidance for conducting interviews effectively. The actual interviews happen outside this workflow -- with real people, in real conversations. This is your reference guide for before, during, and after each session.
+
+## Why This Matters
+
+Even with a perfect script and perfect participants, poor interview technique will produce poor data. The difference between a novice interviewer and a skilled one is not the questions they ask -- it's how they listen, when they probe, and what they notice beyond the words.
+
+## Before Each Interview
+
+### Setup Checklist
+
+- [ ] Review the participant's screening survey answers
+- [ ] Have your interview script printed or visible (but not as a rigid checklist)
+- [ ] Test your recording setup (if recording)
+- [ ] Prepare a note-taking template (participant ID, date, key sections)
+- [ ] Close distracting tabs and notifications
+- [ ] Have water nearby (you'll be talking for 30-45 minutes)
+- [ ] Arrive 2-3 minutes early to the call/meeting
+
+### Mental Preparation
+
+Remind yourself of three things:
+1. **You are here to learn, not to validate.** If the participant contradicts your assumptions, that is the most valuable data.
+2. **Silence is your friend.** When you stop talking, people fill the silence with their real thoughts.
+3. **You are not selling.** Do not pitch, explain, or defend your product. Just listen.
+
+## During the Interview
+
+### The TEDW Framework for Active Listening
+
+Use these four techniques throughout the interview:
+
+**T - Tell me more**
+When a participant says something interesting, don't jump to the next question. Say: "Tell me more about that." This simple phrase produces more insight than any cleverly designed question.
+
+**E - Echo**
+Repeat the participant's last few words as a question. Participant: "It was really frustrating when the report didn't export correctly." You: "Didn't export correctly?" This encourages them to elaborate without leading.
+
+**D - Dig into specifics**
+When answers are vague, ask for concrete details:
+- "Can you give me a specific example?"
+- "When was the last time that happened?"
+- "Walk me through exactly what you did."
+- "How many times did that happen last month?"
+
+**W - Why (but carefully)**
+"Why" questions can feel interrogative. Soften them:
+- Instead of "Why did you do that?" try "What was going through your mind when you decided to...?"
+- Instead of "Why don't you use X?" try "What has kept you from trying X?"
+- "Help me understand your thinking there" works well as an alternative
+
+### Managing Common Interview Situations
+
+**The participant gives one-word answers:**
+- Switch to storytelling prompts: "Walk me through a specific time when..."
+- Share that you're genuinely curious (not testing them)
+- Try "What else?" or "And then what happened?"
+
+**The participant goes off-topic:**
+- Let them finish their thought (sometimes tangents reveal gold)
+- Gently redirect: "That's really interesting. Going back to [topic], I'd love to hear..."
+- Note the tangent -- it might reveal an unmet need you hadn't considered
+
+**The participant asks what YOU think:**
+- Deflect: "I'm really curious about your perspective first."
+- If pressed: "I have some ideas, but I don't want to bias your answers. Can I share after we finish?"
+
+**The participant gets emotional:**
+- Acknowledge: "It sounds like that was really frustrating."
+- Pause. Give them space.
+- Ask: "Would you like to continue, or should we skip to a different topic?"
+- Never minimize: Do NOT say "It's not that bad" or "That's normal"
+
+**The participant wants to design the solution for you:**
+- Listen to the underlying need, not the proposed solution
+- "That's an interesting idea. What problem would that solve for you?"
+- "If that existed, what would be different about your day?"
+
+**You realize your question was poorly worded:**
+- It's fine to rephrase mid-interview: "Let me ask that differently..."
+- Note it for script revision before the next interview
+
+### Note-Taking During the Interview
+
+**What to capture in real-time:**
+- Direct quotes (mark with quotation marks)
+- Emotional moments (mark with [EMOTION: frustration/excitement/surprise])
+- Contradictions between what they say and what they do (mark with [CONTRADICTION])
+- Specific numbers, frequencies, or timeframes they mention
+- Names of tools, processes, or people they reference
+
+**What NOT to do while taking notes:**
+- Don't try to write everything down -- you'll stop listening
+- Don't interpret while writing -- capture raw data now, analyze later
+- Don't type loudly if on a video call -- it signals you're not fully present
+
+**If recording:** Use the recording as backup, not as a replacement for notes. You will not re-listen to 8 hours of recordings. Your in-the-moment notes are more valuable.
+
+## After Each Interview
+
+### Immediate Debrief (within 15 minutes)
+
+Complete this within 15 minutes of ending the interview, while everything is fresh:
+
+1. **Top 3 things that surprised me:**
+   - {surprise 1}
+   - {surprise 2}
+   - {surprise 3}
+
+2. **Strongest quotes:**
+   - "{quote 1}" -- Context: {when they said it}
+   - "{quote 2}" -- Context: {when they said it}
+
+3. **Key behaviors observed:**
+   - {behavior 1} -- Frequency: {how often}
+   - {behavior 2} -- Frequency: {how often}
+
+4. **Assumptions confirmed or challenged:**
+   - CONFIRMED: {assumption} -- Evidence: {what they said/did}
+   - CHALLENGED: {assumption} -- Evidence: {what they said/did}
+
+5. **What would I ask differently next time?**
+   - {adjustment for next interview}
+
+### Script Refinement
+
+After each interview, ask yourself:
+- Did any question consistently produce shallow answers? Rewrite it.
+- Did a follow-up probe reveal something the main question missed? Promote it.
+- Did the participant bring up a topic I hadn't considered? Add a question about it.
+- Was the interview too long or too short? Adjust.
+
+**Important:** It is normal and expected to refine your script between interviews. The first interview is always the roughest. By interview 3-4, your script should be significantly sharper.
+
+---
+
+## Your Turn
+
+Review this guidance and confirm you're ready to begin interviewing. If you have specific concerns about interview technique or situations you expect to encounter, let's discuss them now.
+
+**Tip:** Your first interview will feel awkward. That's normal. The goal is not perfection -- it's genuine curiosity. If you authentically care about understanding the participant's world, the technique will follow.
+
+## Next Step
+
+After you've conducted your interviews (or at least the first 2-3), return and I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/user-interview/steps/step-05-findings.md

package/_bmad/bme/_vortex/workflows/user-interview/steps/step-05-findings.md

@@ -0,0 +1,163 @@
+---
+step: 5
+workflow: user-interview
+title: Capture Findings
+---
+
+# Step 5: Capture Findings
+
+Now that you've conducted interviews, it's time to organize raw findings into a structured format that makes synthesis possible. This step focuses on accurate capture, not interpretation.
+
+## Why This Matters
+
+Raw findings are your primary evidence. If you skip straight to "insights" without properly documenting what you actually heard, you risk:
+- Cherry-picking quotes that confirm your bias
+- Losing important details that only matter later
+- Being unable to trace an insight back to its source
+- Making claims you can't defend ("users want X" -- which users? when? what did they actually say?)
+
+## The Difference Between Findings and Insights
+
+| Findings (this step) | Insights (next step) |
+|----------------------|---------------------|
+| What participants said and did | What it means for your product |
+| Raw, specific, attributed | Synthesized, generalized, thematic |
+| "P3 said she checks Slack 40 times per day" | "Notification anxiety drives compulsive tool-checking" |
+| Observable, verifiable | Interpreted, debatable |
+| Per-participant | Across participants |
+
+**In this step, focus entirely on findings.** Resist the urge to interpret. Capture what happened.
+
+## Your Task
+
+For each interview, create a structured findings capture using the format below:
+
+### Per-Interview Findings Template
+
+```
+## Interview: {Participant ID}
+
+**Date:** {date}
+**Duration:** {minutes}
+**Channel:** {video / phone / in-person}
+**Participant context:** {brief relevant background from screening}
+
+### Key Behaviors Observed
+- {What they actually do, with specifics}
+- {Frequency, tools, workarounds}
+
+### Notable Quotes
+- "{Exact quote}" -- Context: {what prompted this}
+- "{Exact quote}" -- Context: {what prompted this}
+
+### Pain Points Mentioned
+- {Pain point} -- Severity: {how much it bothers them, in their words}
+- {Pain point} -- Severity: {how much it bothers them, in their words}
+
+### Current Workarounds
+- {How they cope with the problem today}
+- {Tools, hacks, manual processes}
+
+### Emotional Moments
+- [EMOTION: {type}] When discussing {topic}: {what happened}
+
+### Contradictions
+- [CONTRADICTION] Said {X} but earlier described doing {Y}
+
+### Unexpected Findings
+- {Anything you didn't anticipate hearing}
+
+### Assumptions Affected
+- CONFIRMED: {assumption} -- Evidence: {quote or behavior}
+- CHALLENGED: {assumption} -- Evidence: {quote or behavior}
+- NEW QUESTION: {something you now want to investigate}
+```
+
+### Quality Standards for Findings
+
+**Each finding should be:**
+
+1. **Specific** -- not "users are frustrated" but "P2 described spending 20 minutes every Monday morning reconciling data between Slack and their spreadsheet, calling it 'the most pointless part of my week'"
+
+2. **Attributed** -- tied to a participant ID so you can trace it back
+
+3. **Behavioral when possible** -- what they do matters more than what they say they'd do
+
+4. **Contextualized** -- include the circumstances that make this finding meaningful
+
+5. **Verbatim when quoting** -- use exact words, not paraphrases (paraphrasing introduces your bias)
+
+### How to Handle Conflicting Findings
+
+When participants contradict each other, capture both sides without resolving the conflict. Note the contradiction explicitly:
+
+> **CONFLICT:** P1 says email notifications are essential for staying on top of tasks. P4 says email notifications are "pure noise" and turned them all off. Both manage teams of similar size.
+>
+> **Possible explanation to explore:** Different management styles? Different team cultures? Different email volumes?
+
+Conflicting findings are not a problem -- they're a signal that your user base is not monolithic. This is valuable.
+
+### Finding Saturation
+
+Track whether you're still learning new things:
+
+| Interview | New findings | Repeated findings | New themes |
+|-----------|-------------|-------------------|------------|
+| P1 | 8 | 0 | 5 |
+| P2 | 6 | 3 | 2 |
+| P3 | 4 | 5 | 1 |
+| P4 | 2 | 7 | 0 |
+| P5 | 1 | 8 | 0 |
+
+When new findings approach zero and repeated findings dominate, you've likely reached saturation. This is a signal that you can move to synthesis.
+
+If you're still hearing completely new things after 6-8 interviews, consider whether your participant criteria are too broad (you may be talking to fundamentally different user types).
+
+## Example
+
+**Interview: P3**
+
+**Date:** 2026-02-15
+**Duration:** 38 minutes
+**Channel:** Video (Zoom)
+**Participant context:** Marketing manager at 15-person agency, uses Trello + spreadsheets, evaluated Asana 3 months ago
+
+### Key Behaviors Observed
+- Checks Trello board 6-8 times per day but only updates cards twice per week
+- Maintains a personal spreadsheet that mirrors the Trello board with additional columns for "real" deadlines vs. "Trello" deadlines
+- Screenshots Trello board for weekly client updates instead of generating reports
+
+### Notable Quotes
+- "Trello is where work goes to look organized. My spreadsheet is where I actually manage things." -- Context: Explaining why she maintains both systems
+- "I tried Asana for two weeks and it felt like I was learning a new language just to do what I already do in Excel." -- Context: When asked about her Asana evaluation
+- "My team would riot if I changed tools again. We switched to Trello eight months ago and some people still complain." -- Context: When asked about barriers to switching
+
+### Pain Points Mentioned
+- Duplicate data entry between Trello and spreadsheet -- Severity: "It's annoying but I've accepted it"
+- Cannot generate client-ready reports from Trello -- Severity: "This is the one that actually costs me time"
+- Team members forget to update cards -- Severity: "I end up chasing people on Slack, which defeats the purpose"
+
+### Emotional Moments
+- [EMOTION: frustration] When describing the screenshot workflow for client reports: voice pitch rose, spoke faster, used the word "ridiculous" twice
+
+### Contradictions
+- [CONTRADICTION] Said "Trello works fine for us" but described 3 significant workarounds and a shadow spreadsheet system
+
+### Assumptions Affected
+- CHALLENGED: "Users find PM tools too complex" -- P3's issue was not complexity but migration cost (learning curve + team resistance)
+- CONFIRMED: Spreadsheets serve needs that PM tools don't -- specifically custom reporting
+- NEW QUESTION: How common are "shadow spreadsheets" alongside PM tools?
+
+---
+
+## Your Turn
+
+Please capture your findings for each completed interview using the format above.
+
+**Tip:** Do this as soon as possible after each interview. Memory fades fast. Findings captured 24 hours later are significantly less accurate than those captured within 30 minutes.
+
+## Next Step
+
+When you've captured findings from all interviews (or enough to see patterns), I'll load:
+
+{project-root}/_bmad/bme/_vortex/workflows/user-interview/steps/step-06-synthesize.md

package/_bmad/bme/_vortex/workflows/user-interview/steps/step-06-synthesize.md

@@ -0,0 +1,171 @@
+---
+step: 6
+workflow: user-interview
+title: Synthesize Insights
+---
+
+# Step 6: Synthesize Insights
+
+Now that you have structured findings from multiple interviews, it's time to look across participants and extract patterns, themes, and actionable insights.
+
+## Why This Matters
+
+Individual interviews tell stories. Synthesis tells you what those stories mean together. Without synthesis, you have a collection of anecdotes. With synthesis, you have evidence-backed insights that can drive product decisions.
+
+## The Synthesis Process
+
+### Phase 1: Theme Identification
+
+Look across all your interview findings and identify recurring themes. A theme is a pattern that appears in 3 or more interviews (or 40%+ of your participants, whichever is higher).
+
+**How to identify themes:**
+
+1. **Lay out all findings visually.** If using physical notes, use sticky notes on a wall. If digital, use a spreadsheet or affinity mapping tool.
+
+2. **Group related findings.** Don't force categories -- let them emerge from the data. Ask: "Which findings are about the same underlying thing?"
+
+3. **Name each group.** A good theme name captures the essence in a phrase:
+   - Weak: "Tools" -- too vague
+   - Strong: "Shadow systems maintained alongside official tools" -- specific and descriptive
+
+4. **Count the evidence.** For each theme, note how many participants contributed to it and what their specific evidence was.
+
+### Phase 2: Insight Formulation
+
+Transform themes into insights. An insight is a non-obvious truth that has implications for what you should build or how you should position it.
+
+**The Insight Formula:**
+
+> **[User segment] [behavior/belief] because [underlying motivation], which means [implication for our product].**
+
+**Examples:**
+
+> Small business owners maintain shadow spreadsheets alongside their PM tools because PM tools don't support custom client reporting, which means a PM tool with built-in customizable reports could eliminate the dual-system workflow.
+
+> Team members resist tool changes not because they dislike new tools but because they've invested effort in learning the current one, which means our onboarding needs to show immediate value within the first session to overcome switching inertia.
+
+**Insight Quality Test:** If your insight would be equally true for any product in any market, it's too generic. Good insights are specific to your users and your context.
+
+### Phase 3: Evidence Mapping
+
+For each insight, build an evidence trail:
+
+```
+### Insight: {insight statement}
+
+**Strength:** Strong / Moderate / Emerging
+**Based on:** {N} of {total} participants
+
+**Evidence:**
+- P1: {specific finding or quote}
+- P3: {specific finding or quote}
+- P5: {specific finding or quote}
+- P7: {specific finding or quote}
+
+**Counter-evidence:**
+- P2: {contradicting finding, if any}
+
+**Confidence level:** High / Medium / Low
+**Why this confidence level:** {explanation}
+```
+
+**Strength ratings:**
+- **Strong:** Appeared in 60%+ of interviews with consistent behavioral evidence
+- **Moderate:** Appeared in 40-60% of interviews or with mixed behavioral evidence
+- **Emerging:** Appeared in 2-3 interviews; interesting but needs more validation
+
+### Phase 4: Implications and Recommendations
+
+For each strong or moderate insight, define what it means for your product:
+
+1. **What should you build, change, or stop doing?**
+   - Be specific. Not "improve onboarding" but "show value within the first 3 minutes by auto-importing existing data"
+
+2. **What assumption was validated or invalidated?**
+   - Reference back to your original research goals from Step 1
+
+3. **What new questions emerged?**
+   - Research always generates more questions. Capture them for the next round.
+
+4. **What is the recommended next action?**
+   - Build a prototype? Run a survey to quantify? Do more interviews with a specific segment? Pivot your approach?
+
+### Phase 5: Research Quality Assessment
+
+Honestly evaluate the quality of your research:
+
+**Sample quality:**
+- Did you talk to the right people? Were any participants clearly outside your target?
+- Was there enough diversity? Or did everyone have the same perspective?
+
+**Question quality:**
+- Did any questions consistently produce shallow answers? (Note for future scripts)
+- Did you discover that a key question was missing?
+
+**Confidence assessment:**
+- Which insights do you feel confident acting on?
+- Which need more evidence before you commit resources?
+- What would you do differently in the next round of interviews?
+
+## Your Task
+
+Complete the synthesis using the structure above:
+
+1. Identify 3-5 themes from your findings
+2. Formulate an insight for each theme (using the insight formula)
+3. Map evidence for each insight
+4. Define implications and next actions
+5. Assess research quality honestly
+
+## Example
+
+**Theme 1: Shadow Systems**
+- 5 of 8 participants maintain a parallel tracking system (spreadsheet, personal notes, or second tool) alongside their "official" PM tool
+- The shadow system always contains information the official tool doesn't support: custom deadlines, client-specific views, personal priority rankings
+
+**Insight:** Small business owners maintain shadow systems alongside PM tools because no single tool supports both team coordination AND client reporting in the format they need, which means a tool that eliminates the need for shadow systems would save 3-5 hours per week of duplicate work.
+
+**Evidence:** P1 (spreadsheet), P3 (spreadsheet), P4 (Notion alongside Trello), P6 (personal task list), P8 (spreadsheet + calendar)
+**Counter-evidence:** P2 uses only Asana but acknowledges "it doesn't do everything I need" and compensates with Slack reminders
+**Confidence:** High -- consistent behavioral evidence across 5 of 8 participants
+**Implication:** Build a prototype with customizable views that serve both team-facing and client-facing needs. Test whether eliminating the shadow system is a compelling enough value proposition.
+
+---
+
+## Your Turn
+
+Please synthesize your findings using the phases above. Take your time -- this is where raw data becomes actionable intelligence.
+
+**Tip:** The best insights often come from contradictions and surprises, not from confirmations. Pay special attention to findings that challenged your assumptions.
+
+## Workflow Complete
+
+After synthesis, I'll generate your complete user interview artifact using the template:
+
+{project-root}/_bmad/bme/_vortex/workflows/user-interview/user-interview.template.md
+
+Your artifact will include:
+- Research goals and context
+- Final interview script (as refined through the process)
+- Participant summary
+- Raw findings per interview
+- Synthesized themes and insights with evidence trails
+- Recommendations and next actions
+- Research quality assessment
+
+---
+
+## Vortex Compass
+
+Based on what you just completed, here are your evidence-driven options:
+
+| If you learned... | Consider next... | Agent | Why |
+|---|---|---|---|
+| Interview insights ready for synthesis | research-convergence | Mila 🔬 | Converge interview findings into problem definition (HC1) |
+| Riskiest insights identified | lean-experiment | Wade 🧪 | Test your riskiest insight with an experiment |
+| Want to synthesize across users | empathy-map | Isla 🔍 | Map patterns across all interview subjects |
+
+> **Note:** These are evidence-based recommendations. You can navigate to any Vortex agent
+> at any time based on your judgment.
+
+**Or run Max's [VN] Vortex Navigation** for a full gap analysis across all streams.