# Expert Pov

> Expert POV

- Skill: `matteotitta/expert-pov` (Agent Skill)
- Install (CLI): `npx skillmds@latest add matteotitta/expert-pov`
- Raw SKILL.md: https://api.skillmd.com/api/skills/matteotitta/expert-pov/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: matteotitta (https://skillmd.com/u/matteotitta)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/matteotitta/expert-pov

---


# Expert POV

Extract and synthesize a founder's unique point of view on their market, craft, and philosophy. Produces the raw material for authority-led positioning — beliefs, stances, and perspectives that differentiate the founder.

**Core philosophy:** Authority comes from having a distinct perspective, not just expertise. Surface what the founder *actually believes* (including contrarian views) and synthesize into OBI candidates.

## When to use

**Invoke when user says:**
- "expert POV / founder POV for [founder/company]"
- "one big idea / OBI for [founder]"
- "authority positioning / thought leadership foundation"
- "contrarian beliefs / what makes [founder] different"
- "founder interview questions"

**Do NOT invoke when:**
- User wants product messaging without founder POV → `product-messaging`
- User wants competitor analysis → `competitor-research`
- User wants ICP research → `icp-behavioural`
- User wants content strategy without POV foundation → run this first

## Inputs

**Required:** Company/founder name; industry/category (for question customization).

**Optional (improves quality):** company context, existing content (LinkedIn posts, podcasts, interviews), competitor positioning, founder background.

**Validation before proceeding:** founder name provided; industry known or discoverable; collection method agreed (async vs live vs content mining). If missing, ask the user; offer to run `company-context` first if needed.

## Process

```
Phase 1 — Question generation (25-30 questions across 6 POV dimensions)
       ↓
Phase 2 — Founder input (async / live / content mining)
       ↓
Phase 3 — POV extraction (beliefs, hot takes, origin stories, taste)
       ↓
Phase 4 — OBI synthesis (cluster → candidates → score → develop top OBI)
       ↓
Review Gate 2 (Deeper Review) [Approve] [Iterate OBI] [Expand themes]
```

**Phase walkthroughs (with checkpoints + outputs):**

**Phase 4 explicitly checks the 100 Posts Test — could the founder genuinely write 100 authentic posts about this OBI?**

## Frameworks and references

| Topic | File | Contents |
|-------|------|----------|
| 6 POV dimensions | the premium reference | Belief, Craft, Market, Story, Taste, Vision — with 5 core questions per dimension |
| Full question bank | the premium reference | Prioritized questions by dimension with follow-ups |
| OBI frameworks | the premium reference | What makes a good OBI, formula patterns, 100 Posts Test, Andy Raskin Strategic Narrative |
| Output template | the premium reference | Canonical 10-section deliverable structure (Exec summary → Belief map → Hot takes → Stories → Taste → Theme clusters → OBI candidates → Recommended OBI → Raw material → Gaps) |
| Worked walkthroughs | the premium reference | Belief extraction + OBI candidate examples + anti-examples |
| Full worked example | `examples/example-matteo-titta.md` | End-to-end "Taste + Systems" run for Matteo Tittarelli / Genesys Growth |
| Quality + guardrails | the premium reference | Anti-hallucination rules, pre-delivery checklist, iteration prompts, skill integration map, MCP data integration |

**Scoring matrix (Phase 4.3):** Authenticity 30% · Differentiation 25% · Memorability 20% · Scalability 15% · Business fit 10%. Score 1-5; weighted total. Full criterion definitions in the premium reference.

## Research substrate (Exa)

Default research substrate per `.claude/rules/exa-protocol.md` (auto-loaded for research, audit, competitor, ICP, AEO, content sourcing work).

**Primary tools for this skill:** `web_search_exa` and the plugin `/search` slash command for parallel-subagent dispatch on founder content mining + POV expansion.

**Tool surface during migration:** prefer plugin namespace `mcp__plugin_exa_exa__web_search_exa` (after `claude plugin i exa@claude-plugins-official`); legacy `mcp__exa__web_search_exa` still mounted; both backends route to the same Exa API.

**Citation:** every Exa-derived claim uses `[VERIFIED: exa_search, {url}, accessed {YYYY-MM-DD}]` per `.claude/rules/ontology.md`.

**Quality gate:** ≥3 sources per major claim, ≥50% `[VERIFIED]` confidence, date filter for any "recent / latest" claim, no `WebSearch` fallback without flagging the data gap.

Worked examples + tool catalog: `.claude/skills/meta-skills/exa/`.