PM - Customer Segments & Personas

Agent mode (--agent)

Supports --agent: runs autonomously in a subagent, drafts the artifact from existing inputs, and returns a short summary + coverage note.

• No flag → interactive (default); if inputs are heavy, offer agent mode.
• --agent → obey. First check inputs are complete. Anything missing: do NOT invent it - mark [ASSUMED - what/why] in the output and summary. Never hallucinate to fill a gap.
• Review required: the artifact contains commitments - after drafting, require the user's review before finalizing; do not close decisions autonomously.

Standalone run

Runs with or without a Pureinn workspace - it needs no other skill's artifact, so it works as a single tool from an install that never ran /pureinn.

• No workspace: create just the folder this skill writes into, or write to a path the user names, and say where the file went. Do not scaffold a project, do not invent a state.json, and do not send the user to /pureinn first.
• Read pureinn-variables.md / state.json where a value is actually used, not at the top of the run. Missing value: continue and name the capability it costs.

What this skill does

Takes raw VOC (Voice of Customer) data - interview transcripts, synthetic interview outputs, survey results, behavioral observations - and produces:

1. Customer Segments (distinct groups with shared needs)
2. Personas (1-2 per key segment)
3. Early Adopters Profile

This is a "bring your data" skill. Claude synthesizes, structures, and formalizes patterns from the input. No invented personas without data.

Run this before jtbd-building - JTBD analysis uses personas as input.

Provenance discipline: every persona claim carries its source class - real VOC (interviews, surveys, observation), [CLIENT-ASSERTED] (a client/sponsor described their own users - input, not evidence), or [ASSUMED] (our inference). In commissioned builds the client's description of their users is the most common input - treat it as a hypothesis to validate with real users, never as validated research.

Three populations (commissioned builds): the client's customers are not the only users. The client's staff (admins, operators, support - they use the product daily) are legitimate personas too, and skipping them is the classic launch failure. If discovery captured staff roles (see Client Discovery meeting notes), produce staff personas alongside customer personas.

Dependencies

Recommended before running:

• pm-project-charter - target customer direction and geography
• pm-market-analysis - segment data provides foundation for persona development

Produces artifacts used by:

• jtbd-building - personas are the input context for JTBD analysis
• pm-problem-validation - personas + segments are Track D input
• design-thinking - personas feed the Define stage
• pm-prd - customer segments and personas are core PRD sections

Step 0: PREREQ check + current state

What this skill needs: Raw VOC data - interview transcripts, synthetic interview outputs (SynthFolk, ChatGPT), survey data, behavioral observations, support tickets.

If you have research data: Paste it or reference the files - proceed to structured synthesis (Path A).

If you have no data at all: Do not block. Immediately offer Path B:

Nemáš výskumné dáta. Môžeme pracovať dvomi spôsobmi:

A) Syntetické persony z assumption-based kontextu
   Vygenerujem persony na základe toho čo viem o tvojom produkte,
   priemysle a cieľovom segmente. Označím ich [ASSUMED - validate with real users].

B) Guided elicitation
   Opýtam ťa 5-7 cielených otázok o tvojich zákazníkoch.
   Výsledok bude slabší než z reálnych interview, ale lepší než nič.

[AskUserQuestion: ktorá cesta? A odporúčaná (rýchlejšia), B (viac kontext)]

Existing artifacts check:

• Do Customer Segments / Personas / Early Adopters Profile already exist? If yes and they look current, surface the skip option: "Persony vyzerajú aktuálne. Preskočiť a pokračovať na /jtbd-building?"
• Does a Market Analysis exist? Cross-reference segments.
• Does a JTBD Analysis exist? Check consistency with jobs and pains.

Gaps check: personas without direct quotes, segments without pain intensity ranking, early adopter profile missing "where to find them" - flag but do not block.

Multi-source handling (Path A with multiple research files): Before synthesizing, assess source alignment per dimension:

• Where sources agree: cite as high-confidence signal ("X sources confirm...")
• Where sources conflict: do NOT force resolution. Surface as: [CONFLICT - [Source A] says X / [Source B] says Y → hypothesis, validate with more data] A conflict between tools (e.g., deep research vs. synthetic interviews) often reflects a real segment split or methodology bias - name it, do not paper over it.

Re-run with new data (delta mode): If existing persona artifacts are present and user is adding new research:

1. State what was in the previous version (key claims per persona/segment)
2. Compare new inputs against previous version
3. Update only what new evidence supports - mark each change: [UPDATED - previous: X / new evidence: Y]
4. Preserve claims that new evidence neither confirms nor contradicts - mark: [UNCHANGED - not addressed in new data]
5. At the end of the artifact, add a Delta Summary: what changed, what was confirmed, what is now in conflict
6. Surface cascade: which downstream artifacts should be re-checked (jtbd-building, pm-problem-validation, PRD personas sections, feature KANO/VxC classifications)

Step 1: Gather inputs

First, ask:

Do you have research data to work with?

A) Yes - I have interviews, surveys, or research notes to paste B) No - I'm working from assumptions and domain knowledge

Path A - Research data available

Use AskUserQuestion tool for all questions at once:

I need VOC data for Customer Segments + Personas.

1. PRODUCT
   What does the product do? For whom? (1-2 sentences)

2. DATA SOURCES (paste what you have available)
   Interview transcripts (direct / synthetic from SynthFolk/ChatGPT):
   [paste here]

   Survey data / responses:
   [paste here]

   Observations / field research:
   [paste here]

   Other insights (forums, reviews, support tickets...):
   [paste here]

   Or point me at files/folders - e.g. the workspace meetings/ folder
   (customer-discovery and client-discovery notes). I read them in full,
   recursively. Client claims about users from client-discovery notes
   enter as [CLIENT-ASSERTED], not as VOC.

3. CONTEXT
   How many interviews / respondents do you have in total?
   What types of people did you talk to? (role, industry, experience level...)
   What were the key questions / topics in the interviews?

4. HYPOTHESES
   What segments are you already suspecting? (if you have a preliminary view)
   Who do you think is the ideal early adopter and why?

Path B - No research data (guided elicitation)

No research data yet. Guide the user through 3 rounds of questions to extract maximum knowledge. The goal is to replace interview and observation work with structured founder knowledge. Output will be marked as assumption-based.

Path B - specific downstream risks by assumption type

• Who the customer is (role, context, org size) assumed: in pm-hypotheses Plan mode, experiment recruitment criteria will be based on this profile. If the profile is off, you will be interviewing the wrong people and generating false positive validation signals.
• How they discover and adopt tools assumed: this feeds directly into the Channel block of Lean Canvas and acquisition stage of AARRR (KPIs). A wrong adoption behavior assumption means your go-to-market channel choice is built on an untested premise.
• Pain description from founder perspective, not customer words: pm-features-list uses the persona to classify features as Must-be, Performance, or Delighter. A feature that feels Must-be from your perspective may be Indifferent to the actual customer - which affects MVP scope decisions.
• Early adopter profile assumed: the first 10-50 customers matter disproportionately. If the assumed early adopter profile does not match who actually responds to your outreach, you will spend acquisition budget targeting the wrong person.

One real interview with someone matching this profile - before Phase 3a - closes the largest share of this risk at minimal cost.

Group 1 of 3 - Who is this person?

Use AskUserQuestion tool for these two questions:

What is their organization or context size?

A) Solo / freelancer - works independently B) Small team - 2 to 15 people C) Mid-size company - 16 to 200 people D) Large enterprise - 200+ people

How tech-savvy are they compared to average?

A) Power user - adopts new tools early, comfortable with change B) Average - uses tools that are widely adopted and proven C) Reluctant - resists change, needs strong reason and support D) Varies significantly across the segment

Then ask these as plain text:

What is their role or job title? What are they responsible for day-to-day?

What are they measured on? What does a good week look like for them?

After receiving answers, show summary and ask: "Is this the right person? Anything to adjust before we go deeper?"

Group 2 of 3 - The problem

Ask all four as plain text:

Describe the problem from their perspective - not what your product does, but what pain they feel and when.

What do they currently use to deal with this? Name specific tools, workarounds, or manual processes.

What does this problem cost them? Be specific - time wasted per week, money lost, stress level, risk, missed revenue.

When does the pain hit hardest? What specific situation or moment triggers the need for a solution?

After receiving answers, show summary and ask: "Does this match what you've observed or heard? Anything to correct?"

Group 3 of 3 - Behavior and adoption

Use AskUserQuestion tool for these two questions:

How do people like this typically discover and adopt new tools?

A) Actively seek new solutions - look for better ways on their own B) Peer-driven - adopt when colleagues or trusted contacts recommend C) Top-down - adopt when their company mandates it D) Pain-driven - only switch when current solution breaks badly enough

Is your understanding of this person based on:

A) Direct conversations with real people matching this profile B) Indirect observation (forums, reviews, support tickets, secondhand) C) Domain expertise and pattern recognition - you know this world well D) Mostly assumptions - not yet validated with real people

Then ask these as plain text:

Why haven't they solved this already? What's blocking them from fixing it today?

What would make them try your product immediately - and what would make them hesitate?

Do you know a specific real person who fits this profile? Describe what makes them the ideal early adopter.

After receiving answers, show a complete summary of everything collected across all 3 groups. Ask for final confirmation before proceeding to artifact generation.

Note at the top of every generated artifact: > Assumption-based - built from founder knowledge, not validated research. Treat as hypotheses to be tested.

Step 2: Generate artifacts

Generate in English.

Segmentation quality gate (apply before writing the Segment Map):

• A segment is valid only if it changes a decision - a different need, buying behavior, or usage context. If two "segments" would get the same product, the same price, and the same message, they are one segment: merge them and say so.
• Demographic labels (company size, industry, geography) are proxies, not segments. Name the underlying need or behavior difference behind the label, or drop the split.
• For MVP work, cap at 2-3 personas. A fourth persona almost always signals unwillingness to choose - force-rank instead and park the rest as [LATER].

ARTIFACT 1: Customer Segments

# Customer Segments - [Product Name]

> **Phase:** 2 - Discovery (Track D: VOC - Customer Discovery)
> **Date:** [date]
> **Data basis:** [X interviews / Y survey responses / Z synthetic interviews]

---

## Segmentation Overview

[1-2 sentences: how we segmented - what primary dimension drove the segmentation (behavior, need, role, context)]

---

## Segment Map

| Segment | Size (est.) | Primary Need | Key Differentiator from Other Segments |
|---|---|---|---|
| [Segment 1 name] | [X% of TAM / X,000 users] | [core job to be done] | [what makes them distinct] |
| [Segment 2 name] | | | |
| [Segment 3 name] | | | |

---

## Segment Profiles

### Segment 1: [Name]

**Who they are:** [2-3 sentences - role, context, scale, typical situation]

**Primary goal:** [What they're trying to achieve]

**Current solution:** [How they solve the problem today]

**Key frustrations with current solution:**
- [Frustration 1]
- [Frustration 2]

**What they'd pay for:** [what value they recognize and would buy]

**Revenue potential:** [size x ARPU estimate]

**Strategic priority:** High / Medium / Low - [one-line reason]

---

### Segment 2: [Name]

[same structure]

---

### Segment 3: [Name]

[same structure]

---

## Segment Prioritization

| Segment | Market Size | Pain Intensity | Willingness to Pay | Accessibility | Priority |
|---|---|---|---|---|---|
| [Segment 1] | H/M/L | H/M/L | H/M/L | H/M/L | **1** |
| [Segment 2] | H/M/L | H/M/L | H/M/L | H/M/L | **2** |
| [Segment 3] | H/M/L | H/M/L | H/M/L | H/M/L | **3** |

**Primary segment for MVP:** [Segment X] - [brief reason]

ARTIFACT 2: Personas

# Personas - [Product Name]

> **Phase:** 2 - Discovery (Track D: VOC - Customer Discovery)
> **Date:** [date]
> **Note:** Personas are composites based on research data, not fictional characters.

---

## Persona 1: [Name] ([Segment])

**Role:** [Job title / context]
**Age range:** [X-Y]
**Context:** [Brief description - company size, geography, experience level]

### Goals

| Goal | Priority |
|---|---|
| [Primary goal - what they're trying to achieve] | High |
| [Secondary goal] | Med |
| [Tertiary goal] | Low |

### Pains

| Pain | Intensity | Current Workaround |
|---|---|---|
| [Primary pain] | High / Med / Low | [what they do today] |
| [Secondary pain] | | |
| [Third pain] | | |

### Behaviors & Habits

- [Relevant behavior 1: e.g., Manages everything via spreadsheets]
- [Relevant behavior 2]
- [Relevant behavior 3]

### Tools They Use Today

[List tools relevant to the problem space]

### Quotes (from research)

> "[Direct quote illustrating their main pain]"

> "[Direct quote illustrating what they want]"

### Empathy Map

| Quadrant | Findings |
|---|---|
| **Says** | [direct quotes, public statements] |
| **Thinks** | [internal thoughts, hopes, fears - often contradicts what they say] |
| **Does** | [observed behaviors, workarounds, actual usage] |
| **Feels** | [emotional states around the problem - frustration, anxiety, pride] |
| **Sees** | [their environment - what competitors' users have, what peers do] |
| **Hears** | [influencers, colleagues, reviews - who shapes their opinion] |

[Fill from research data; mark quadrants without evidence `[TODO - insufficient input]` rather than inventing. For the primary persona this map is required; for secondary personas optional.]

### Current Journey (as-is)

[Step-by-step: how they handle the problem today - where they lose time, money, or energy, and the emotional low points. 3-6 steps, from research. This seeds pm-process-flows in Phase 4.]

### What Success Looks Like for Them

[1-2 sentences: what does their life look like when this problem is solved]

### Why They'd Adopt Our Product

[1-2 sentences: the trigger + the value that converts them]

### Why They Might Not

[Main objection / friction / risk from their perspective]

---

## Persona 2: [Name] ([Segment])

[Same structure]

---

## Persona Usage Guide

| Persona | Use for | Don't use for |
|---|---|---|
| [Persona 1] | [MVP feature prioritization, onboarding design] | [Enterprise sales scenarios] |
| [Persona 2] | | |

ARTIFACT 3: Early Adopters Profile

# Early Adopters Profile - [Product Name]

> **Phase:** 2 - Discovery (Track D: VOC - Customer Discovery)
> **Date:** [date]

---

## Definition

Early adopters are the first 10-50 customers. They are not average users - they are willing to use an imperfect product because the pain is intense enough.

---

## Early Adopter Profile

**Segment:** [Which segment they come from]
**Persona:** [Which persona best represents them, or custom description]

### Characteristics

| Characteristic | Description |
|---|---|
| Pain intensity | [Extremely high - they can't wait for a perfect solution] |
| Current situation | [e.g., Using 3-4 duct-taped tools, losing X hours/week] |
| Technical comfort | [High / Med - early adopters accept rough edges] |
| Decision speed | [Fast - can decide without lengthy procurement] |
| Feedback willingness | [High - actively want to shape the product] |
| Referral potential | [High - if they love it, they tell others in their network] |

### What Triggers Them to Try a New Solution

- [Trigger 1: e.g., Their current tool raised prices / shut down]
- [Trigger 2: e.g., They just hit a scaling wall]
- [Trigger 3: e.g., A peer recommended something new]

### What They Need from Us (Day 1)

**Must have:**
- [The single thing that makes it worth switching]
- [The table-stakes that can't be missing]

**Nice to have (but OK to lack at launch):**
- [Feature that matters but doesn't block adoption]

**What they'll tolerate:**
- [Rough edges they'll accept if core value is there]

### Where to Find Them

| Channel | Method | Est. Volume |
|---|---|---|
| [e.g., Slack communities] | [e.g., STR host communities in SK/CZ] | [X] |
| [e.g., LinkedIn] | [e.g., Search by title + geography] | [X] |
| [e.g., Referrals from interviewees] | [ask existing contacts] | [X] |

### Success Metric for Early Adopter Phase

**Target:** [X paying early adopters by [date]]
**Signal of success:** [e.g., Retention > 80% after 30 days, NPS > 8]

Internal completeness checklist

Customer Segments must cover:

• Segmentation criteria defined (what dimension drove the split: behavior, need, role, context)
• Each segment: size estimate, primary need, key differentiator from other segments
• Pain intensity per segment compared (which segment hurts most)
• Revenue potential per segment estimated
• Primary segment for MVP selected with rationale
• For B2B SaaS: buyer vs. user distinction per segment (who signs, who uses)

Personas must cover:

• Goals: primary, secondary (what they're trying to achieve)
• Pains: with intensity and current workaround
• Behaviors and habits relevant to the problem space
• Tools they use today (integration expectations)
• Direct quotes from research (not invented)
• What success looks like for them (their definition)
• Why they would adopt / why they might not
• For SaaS/AI: technical sophistication level (affects onboarding design)
• For SaaS/AI: AI adoption readiness and trust level
• For B2B: procurement process and budget ownership (self-serve vs. sales-led)

Empathy Map must cover (per key persona):

• Says - direct quotes
• Thinks - internal thoughts, hopes, fears
• Does - observed behaviors
• Feels - emotional states
• Pains - frustrations, obstacles, risks
• Gains - wants, needs, success measures

Current Journey (as-is) must cover:

• Current process mapped step by step
• Biggest pain points in the current journey identified
• Emotional states at each stage captured
• Where customers lose time, money, or energy

Early Adopters Profile must cover:

• Characteristics that make them early adopters (pain intensity, tech comfort, decision speed)
• Triggers that would make them try a new solution
• Day 1 must-haves vs. what they'll tolerate
• Where to find them (specific channels, communities)
• Success metric for early adopter phase (target number, retention signal)

Notion

Read pureinn-variables.md key "Customer Personas" → if URL present, remind user after saving: Personas saved locally. Update Notion: [Customer Personas URL]

Read pureinn-variables.md key "Customer Interviews" → if DB URL present, check state.json notion_ids.customer_interviews. This DB can be used to log interview notes alongside persona generation. Remind user: Customer interviews can be logged to Notion DB: [Customer Interviews URL]

Save to

pureinn-workspace/[project-slug]/artifacts/phase-2-discovery/customer-segments.md
pureinn-workspace/[project-slug]/artifacts/phase-2-discovery/personas.md
pureinn-workspace/[project-slug]/artifacts/phase-2-discovery/early-adopters-profile.md

Handoff

Čo si teraz má: Segmenty, persony a early adopter profil. JTBD analysis môže začať.

Ďalší krok: /jtbd-building — analyzuje Jobs-to-be-Done a Forces diagram pre každú personu. Alebo spusti /pureinn pre Phase 2 gate check ak sú hotové aj ostatné Track A-C.

Môžeš preskočiť /jtbd-building ak: Máš jasné JTBD insights z rozhovorov a nepotrebuješ formálny Forces diagram pre fázu 3a.