Market Intelligence — Deep Research with Evidence Architecture

TL;DR: Full-spectrum market intelligence on any entity — company, person, territory, or sector — delivered as a branded JM Labs HTML report with evidence chips.

Principio Rector: Evidence over assumption. Every claim carries an evidence tag. No speculation presented as fact.

---

When to Activate

Activate when:

• User says "market intelligence", "market research", "geographic analysis", "sector analysis"
• User says "competitive landscape", "research X company", "profile on X person"
• User says "contact points for", "entry strategy for [market]", "who are the players in [sector]"
• User needs a structured evidence-backed report on a market, company, or public figure [EXPLICIT]

Do NOT activate when:

• User needs a quick one-liner fact (use WebSearch directly instead)
• User is asking about internal company data they own (use input-analysis skill)
• Request is purely financial modeling (use cost-estimation skill)

---

S1 — Input Classification

Classify the target entity before starting research. This determines the research path and data sources. [EXPLICIT]

Entity Type	Definition	Primary Sources
Company	Legal entity, product or service org	LinkedIn, Crunchbase, official site, news
Person	Executive, politician, influencer, decision maker	LinkedIn, Twitter/X, press, company bios
Territory	City, region, country, metro area	Government data, chamber of commerce, World Bank
Sector	Industry vertical, market segment	Gartner, IBISWorld, Statista, trade associations

Classification Protocol:

1. Parse user input for entity name and type [EXPLICIT]
2. If ambiguous (e.g., "Amazon" — company or territory?), ask one clarifying question [INFERRED]
3. If entity has multiple dimensions (e.g., a sector in a territory), run both tracks [INFERRED]
4. Log classification decision with rationale before proceeding [OPEN]

---

S2 — Research Framework

2a. Company Research Track

Layer 1 — Company DNA

• Full legal name, founding year, HQ location [EXPLICIT]
• Business model (B2B/B2C/B2B2C, product/service/platform) [INFERRED]
• Revenue signals (public data, funding rounds, employee count proxies) [INFERRED]
• Ownership structure (private, VC-backed, PE, public, family) [EXPLICIT]

Layer 2 — Competitive Position

• Direct competitors (same ICP, same solution category) [INFERRED]
• Indirect competitors (different approach, same problem) [INFERRED]
• Market share signals (web traffic, app store rankings, media coverage volume) [INFERRED]
• Differentiation claims (website messaging, PR positioning) [EXPLICIT]

Layer 3 — Tech Ecosystem

• Known tech stack (BuiltWith, job postings analysis, GitHub signals) [INFERRED]
• Integrations and partnerships declared publicly [EXPLICIT]
• AI/automation adoption signals [INFERRED]

Layer 4 — Momentum Signals

• Recent funding rounds (last 24 months) [EXPLICIT]
• Key hires and departures (LinkedIn signals) [INFERRED]
• Product launches, pivots, press releases [EXPLICIT]
• Customer wins, case studies, testimonials [EXPLICIT]

2b. Territory Research Track

Demographics Layer

• Population, age distribution, urbanization rate [EXPLICIT]
• Education levels, literacy, language landscape [EXPLICIT]
• Income distribution, middle class size, purchasing power parity [EXPLICIT]

Economic Layer

• GDP, GDP per capita, growth rate (last 3 years) [EXPLICIT]
• Key industries by GDP contribution [EXPLICIT]
• Unemployment rate, informal economy size [EXPLICIT]
• FDI attractiveness, ease of doing business ranking [EXPLICIT]

Tech Ecosystem Layer

• Internet penetration rate, mobile-first vs desktop [EXPLICIT]
• Startup ecosystem maturity (unicorns, accelerators, VCs) [INFERRED]
• Digital payment adoption (fintech penetration) [INFERRED]
• Government digitalization index [EXPLICIT]

Regulatory Layer

• Data protection laws (GDPR equivalent, LGPD, etc.) [EXPLICIT]
• Industry-specific regulations relevant to user's sector [INFERRED]
• Tax regime highlights for business entry [EXPLICIT]

2c. Sector Research Track

• Market size (TAM, SAM, SOM) with source and date [EXPLICIT]
• Growth rate (CAGR, last 5-year trend) [EXPLICIT]
• Key players and concentration (HHI proxy) [INFERRED]
• Technology disruption vectors in the sector [INFERRED]
• Buyer behavior patterns (procurement cycles, decision makers) [INFERRED]
• Regulatory tailwinds and headwinds [EXPLICIT]

---

S3 — Public Figure Profile

Use when entity type = Person (executive, politician, investor, influencer, SME). [EXPLICIT]

Identity Map

• Full name, current role and organization [EXPLICIT]
• Career trajectory (last 3 positions minimum) [EXPLICIT]
• Education credentials (university, field) [EXPLICIT]
• Geographic base and mobility signals [INFERRED]

Social Presence

• LinkedIn URL, connection count proxy, posting frequency [EXPLICIT]
• Twitter/X handle, follower count, engagement style [EXPLICIT]
• Public speaking: conferences, podcasts, panels [EXPLICIT]
• Published content: articles, whitepapers, books [EXPLICIT]

Credibility Map

• Professional recognition (awards, board memberships, advisory roles) [EXPLICIT]
• Media coverage volume and tone [INFERRED]
• Controversies or public criticism (if publicly documented) [EXPLICIT]
• Thought leadership topics (recurring themes in public statements) [INFERRED]

Influence Map

• Who follows/endorses them publicly [INFERRED]
• Organizational networks (alumni, industry groups) [INFERRED]
• Investment portfolio or board seats (if public) [EXPLICIT]

---

S4 — Commercial Contact Points

Extract actionable contact intelligence from public sources only. [EXPLICIT]

Email Pattern Detection

• Infer corporate email pattern from known public emails or breach data indices [INFERRED]
• Common patterns: firstname@domain, f.lastname@domain, firstname.lastname@domain
• Never fabricate — mark as [INFERRED] with confidence level (High/Medium/Low)

LinkedIn Channel

• Direct InMail eligibility (premium connection) [INFERRED]
• Mutual connections that could facilitate warm intro [INFERRED]
• Group memberships for indirect engagement [EXPLICIT]

Direct Channels

• Company contact forms, sales inquiry pages [EXPLICIT]
• Event attendance (upcoming conferences, webinars) [EXPLICIT]
• Press/PR contact (for media-adjacent outreach) [EXPLICIT]

Gatekeeper Mapping

• EA/PA names if publicly visible (LinkedIn, company site) [EXPLICIT]
• Receptionist/front-desk contacts for physical outreach [EXPLICIT]
• Junior team members who manage comms for decision makers [INFERRED]

---

S5 — Output — Branded HTML Report

Generate a self-contained HTML file using JM Labs brand system. [EXPLICIT]

Brand Tokens to Apply:

:root {
  --jml-navy:  #122562;
  --jml-gold:  #FFD700;
  --jml-blue:  #137DC5;
  --jml-bg:    #0d1b3e;
  --jml-surface: rgba(255,255,255,0.05);
  --font-head: 'Poppins', sans-serif;   /* 600-900 weight */
  --font-body: 'Inter', sans-serif;     /* 400-800 weight */
}

HTML Structure:

<div class="jml-doc">
  <header class="slide">   <!-- Hero with title, subtitle, date -->
  <section class="bridge"> <!-- Executive summary / key findings -->
  <section class="mv">     <!-- Main content sections (S1-S4 findings) -->
  <section class="ml">     <!-- Evidence log / sources -->
  <footer class="cta">     <!-- Next steps / recommended actions -->
</div>

Evidence Chips Pattern:

<span class="src explicit">[EXPLICIT]</span>
<span class="src inferred">[INFERRED]</span>
<span class="src open">[OPEN]</span>

Bilingual Support (ES/EN):

<span data-l="es">Inteligencia de Mercado</span>
<span data-l="en">Market Intelligence</span>

JavaScript toggle between data-l="es" and data-l="en" via .gold button.

Report Sections:

1. Executive Summary (key findings, 3-5 bullets)
2. Entity Profile (classification + core facts)
3. Market Position / Competitive Landscape
4. Contact Intelligence (S4 output)
5. Opportunity Map (where to engage, why, how)
6. Evidence Log (all sources with URLs and access date)
7. Recommended Next Steps

---

S6 — Evidence Tagging Protocol

Apply evidence tags to every factual claim in all outputs. No exceptions. [EXPLICIT]

Tag	Meaning	Usage Rule
[EXPLICIT]	Directly stated by a primary source	Use for facts from official sites, filings, press releases
[INFERRED]	Derived logically from available signals	Use for estimates, pattern-based conclusions
[OPEN]	Unknown — requires further research	Use when data gap exists but matters to the user

Tagging Discipline:

• Untagged claims are forbidden in MOAT-grade outputs [EXPLICIT]
• If a claim cannot be tagged, it must be removed or converted to a question [INFERRED]
• [OPEN] tags must include a note on how to resolve them [EXPLICIT]

---

Trade-off Matrix

Dimension	Quick (1h)	Standard (4h)	Deep (8h+)
Source depth	Top 10 web results	30+ sources, cross-validated	Full OSINT + database sweep
Person profile	LinkedIn + basic bio	+ Social + media coverage	+ Network map + influence score
Contact intel	Email pattern only	+ LinkedIn + gatekeeper	+ Event calendar + warm intro path
Output format	Markdown summary	Branded HTML report	HTML + CSV + exec briefing
Confidence level	Medium	High	Very High

---

Assumptions & Limits

• Public data only — no hacking, no social engineering, no dark web [EXPLICIT]
• Email patterns are inferred, never verified without testing [EXPLICIT]
• Territory data accuracy depends on recency of government data sources [INFERRED]
• Sector size figures (TAM/SAM/SOM) must cite source and vintage year [EXPLICIT]
• Person profiles respect privacy: only publicly available information [EXPLICIT]
• Real-time web search required for current signals; static knowledge has cutoff [EXPLICIT]

---

Edge Cases

Scenario	Handling
Entity has same name as another (e.g., "Amazon" river vs. company)	Clarify with one question before starting; log disambiguation decision
Company is stealth / pre-launch with minimal public presence	Flag as [OPEN], deliver what exists, list data gaps explicitly
Public figure has scrubbed their online presence	Document what was found, note the scrubbing as signal, do not speculate
Territory has unreliable government data (conflict zone, sanctions)	Use World Bank / IMF as primary; flag reliability concern
User requests information on a minor (person under 18)	Refuse the person profile; redirect to the organization they're associated with

---

Good vs Bad Example

BAD (untagged, vague):

"Company X is growing fast and has about 500 employees. They use Salesforce."

GOOD (tagged, sourced, specific):

"Company X had 487 employees as of Q3 2024 per LinkedIn headcount [EXPLICIT]. Revenue estimated at $40-60M based on Series B raise of $15M (2023) and SaaS multiples [INFERRED]. Salesforce usage confirmed via BuiltWith scan [EXPLICIT]. Hiring 12 open roles in Engineering as of research date [EXPLICIT]."

---

Validation Gate

• Entity type classified (company/person/territory/sector) with rationale
• Research framework section(s) executed matching entity type
• All factual claims carry evidence tags ([EXPLICIT] / [INFERRED] / [OPEN])
• [OPEN] items include resolution path
• Contact intelligence section completed (even if partial)
• HTML report uses JM Labs brand tokens (navy #122562, gold #FFD700, blue #137DC5)
• HTML report is self-contained (fonts via Google Fonts CDN, no external CSS dependencies)
• Bilingual ES/EN toggle implemented in HTML output
• Evidence log section present with source URLs and access dates
• Trade-off level declared (Quick / Standard / Deep) and scope honored

---

Reference Files

• knowledge/body-of-knowledge.md — Market research methodology foundations
• knowledge/knowledge-graph.md — Entity relationship map for research paths
• references/research-framework.md — OSINT sources, B2B databases, signal detection methods
• evals/ — Test cases with known entities for validation

---

Related Skills

• client-dossier — Deeper company + contacts dossier (next step after market-intelligence)
• client-prospecting — Systematic lead generation using market intelligence output
• b2b-outreach — Convert contact intelligence into outreach sequences
• competitive-positioning — Strategic positioning informed by competitive landscape research