MCP Workflow Guide for Product Managers

The Product Manager Research Workflow

Product managers spend hours context-switching between research tools, spreadsheets, and slide decks. With Koji MCP and Claude, you can run your entire research workflow in one conversation — from hypothesis to stakeholder-ready report.

This guide walks through a real-world scenario: validating whether enterprise users need SSO before you build it.

---

Step 1: Define Your Research Question

Start a new Claude conversation and describe your research goal:

"I am a product manager at a B2B SaaS company. We keep getting SSO requests from enterprise prospects but I am not sure if it is actually blocking deals or just a nice-to-have. Create a customer discovery study to figure this out."

Claude will use koji_create_study with the Customer Discovery methodology to create a study with:

• A focused problem statement about SSO as a deal-blocker
• Key interview questions that avoid leading the witness
• Guardrails based on methodology best practices (e.g., asking about past behavior, not hypotheticals)
• An interview plan with probe points for deeper exploration

Pro tip: Mention your methodology preference if you have one. "Use the Mom Test approach" or "I want a Jobs-to-be-Done framework" will shape the entire interview structure.

---

Step 2: Add Structured Questions for Quantitative Data

This is where MCP gets powerful for PMs. You can mix qualitative depth with quantitative metrics:

"Add these structured questions to my study: an NPS question on a 0-10 scale with anchor probing, a single-choice question asking how they currently handle SSO (options: shared passwords, password manager, manual provisioning, existing SSO, other), and a ranking question asking them to prioritize security, ease of onboarding, compliance, and cost."

Claude uses koji_update_brief to add structured questions with:

• Scale questions — NPS or satisfaction ratings with configurable ranges and labeled endpoints
• Choice questions — Segment respondents by their current behavior or preferences
• Ranking questions — Understand relative priority of features or pain points
• Anchor probing — After a rating, the AI asks "You said X — what would need to change for that to be higher?"

These questions appear as interactive widgets during the interview while the AI interviewer still conducts a natural conversation around them.

---

Step 3: Review and Refine the Brief

After creation, ask Claude to show you the full brief:

"Show me the research brief, interview plan, and all structured questions"

Review the questions. You might refine them:

"The questions look good but add one about what they are currently doing instead of SSO — like password managers or shared credentials. Also increase the probing depth to 2 follow-ups for the NPS question."

Claude will update the study with your refinements using koji_update_brief.

---

Step 4: Publish and Distribute

When you are happy with the brief:

"Publish the study and configure it with voice enabled as default, set the URL slug to 'sso-enterprise-research', and set the headline to 'Help Us Build Better Enterprise Security'"

Claude publishes the study with koji_publish_study and configures branding with koji_configure_study — all in one conversation turn.

For targeted outreach:

"Import these contacts for the study: Sarah Chen (VP Engineering, Acme Corp), James Park (CTO, Betaworks), Maria Rodriguez (IT Director, CloudFirst)"

Each person gets a personalized interview link you can include in email outreach.

---

Step 5: Monitor Progress

Check in on interview completion:

"How are my SSO interviews going? Show me the completed ones."

Claude shows you interview summaries with sentiment and themes — no need to read every transcript. For deeper exploration:

"The interview with Sarah Chen sounds interesting. Show me her full transcript and structured answers."

Claude shows the transcript along with her NPS score, choice selections, and ranking — all with qualitative context from the conversation.

---

Step 6: Analyze Patterns with Quantitative Data

Once you have 5 or more completed interviews, ask for structured analysis:

"Get the study data and tell me what patterns you see. What is the average NPS score? How do respondents rank the security priorities? Is SSO actually blocking deals?"

Claude pulls the aggregated data — per-question statistics, themes, sentiment distribution — and synthesizes findings across all interviews. For structured questions, you get:

• NPS averages and distributions — "Average NPS is 6.2, with 40% detractors"
• Choice frequency counts — "8 of 12 respondents use shared passwords as their current workaround"
• Ranking positions — "Compliance ranked #1 on average, followed by ease of onboarding"
• Cross-referencing — "Respondents who gave NPS below 5 all ranked compliance as their top priority"

---

Step 7: Generate the Report

"Generate a research report and publish it so I can share with stakeholders"

Claude generates a comprehensive report with:

• Executive summary
• Key takeaways ranked by importance
• Theme analysis with supporting quotes
• Charts showing NPS distribution, choice breakdowns, and ranking results
• Recommendations tied to evidence
• Sentiment breakdown

Then publishes it with a shareable link you can drop into Slack, email, or your product brief.

---

Step 8: Export for Deeper Analysis

If you need to bring data into another tool:

"Export the full study data including transcripts and respondent info"

You get structured JSON you can import into Notion, Airtable, or your product analytics stack.

---

The Full Conversation Flow

Here is how the entire workflow looks in a single Claude conversation:

You: Create a customer discovery study about whether SSO is blocking enterprise deals
You: Add an NPS question (0-10), a choice question about current auth methods, and a ranking of security priorities
You: Review the brief and adjust probing depth
You: Publish it with voice enabled, custom URL, and branded headline
You: Import my list of 15 enterprise contacts
[Wait for interviews to complete]
You: Show me the NPS average and feature ranking results
You: Generate a report and publish it for the team
You: Export everything to JSON for our product brief

Eight messages. One conversation. A complete research cycle with both qualitative and quantitative data.

---

PM-Specific Tips

Using Structured Questions for Feature Prioritization

Create a study specifically for feature prioritization:

"Create a study with a ranking question asking users to prioritize these features: SSO, audit logs, role-based access, API access, and custom integrations. Add a scale question for each asking how critical it is (1-5). Use the discovery methodology."

This gives you both relative rankings and absolute importance scores — perfect for roadmap decisions.

Choosing the Right Methodology

Handling Stakeholder Questions

When your VP asks "do users actually want this?", you can respond in real-time:

"What is the average NPS score and top themes from my onboarding study?"

No scrambling for slides. Claude gives you data-backed answers instantly with both quantitative metrics and qualitative context.

Weekly Research Cadence

Set up a recurring workflow:

1. Monday: Check interview progress across active studies
2. Wednesday: Review new completed interviews and structured question results
3. Friday: Generate updated reports if you have new interviews

---

Next Steps

• Tool Reference — Deep dive into each tool
• Structured Questions Guide — Configure question types, scales, and probing
• Best Practices — Tips for better research through MCP
• MCP Overview — Full integration overview

Further reading on the blog

• B2B Customer Research: The Complete Guide for Product Teams (2026) — B2B customer research is harder than B2C — you are navigating buying groups of 10+ stakeholders, gatekeepers, and enterprise procurement cyc
• Best Product Discovery Tools in 2026: The Complete Buyer's Guide — Product discovery is no longer a one-off pre-launch phase — it is a continuous loop. The best teams in 2026 are running weekly customer inte
• Concept Testing: The Complete Guide for Product Teams (2026) — Concept testing validates whether your idea is worth building before you build it. This guide covers methods, question templates, analysis a

<!-- further-reading:blog -->