B2B Research Agent Skill

Research and identify business-to-business engagement opportunities with a problem-first approach. This skill conducts structured web research to find problems worth solving, validate qualification, map buying committees, and produce intelligence that earns the right to a diagnostic conversation.

Core Philosophy: Earn the Right

This skill is built on the principle that progression is earned, not claimed. Research should honestly assess:

Is there a real problem we can solve? — Not assumed, but evidenced
Do we have access to authority? — Specific people, not vague roles
Is there willingness to engage? — Signals of openness, not assumptions

The goal is not to "sell" — it's to qualify whether a conversation is worth having, and if so, to earn the right to have it.

BEAM Integration

This skill produces output designed to feed directly into the marcov.BEAM (Bayesian Evidence-Advancing Markov) selling framework. Every dossier includes:

BEAM Qualification Readiness: Pre-assessment of Stage 1 gate criteria
Problem Domain Analysis: Evidence-based articulation of problems worth solving
Buying Committee Mapping: Stakeholders by BEAM buying role
Qualification Signals: Evidence supporting each Stage 1 gate

BEAM Stage 1 Gates (What Research Must Support)

Gate	Research Must Provide
Problem domain identified	Specific, evidenced problems — not assumptions or generic categories
Access to authority	Named individuals with buying roles — not "someone in IT"
Willingness to diagnose	Signals that suggest openness to a conversation — or honest assessment of barriers

Qualifying out is a valid outcome. If research reveals no real problem or no path to authority, that's valuable intelligence — don't force-fit an opportunity that doesn't exist.

Overview

This skill helps you:

Problem Discovery: Find specific, evidenced problems the prospect is experiencing — the foundation for any engagement
Qualification Assessment: Honestly evaluate whether this prospect warrants pursuit under BEAM criteria
Buying Committee Mapping: Identify stakeholders by BEAM buying role (Economic Buyer, Technical Evaluator, Champion, Gatekeeper)
Access Path Analysis: Map routes to authority — warm introductions, events, content hooks
Engagement Readiness: Assess signals of willingness to have a diagnostic conversation
Competitive Intelligence: Understand what solutions prospects currently use and where gaps create problems
Pipeline Building: Produce qualified prospect lists ranked by fit, timing, and qualification strength

Input

This skill accepts a target company name as its primary input. The user provides the company they want to research, and the skill produces a full engagement dossier and HTML report for that company.

Invocation Examples

/b2b-research-agent GeelongPort
/b2b-research-agent Research BHP for asset management consulting opportunities
/b2b-research-agent Analyse Yarra Trams as a potential client for edge AI

When a target company is provided, skip straight to the focused discovery questions below — do not ask broad questions about target market, ICP, or how many prospects they need. The target is already known.

Discovery Process (CRITICAL)

Before conducting any research, you MUST conduct a brief discovery interview to understand the engagement context and what problem you're trying to solve for them. Adapt the questions based on what was provided in the invocation.

When a Target Company Is Provided (Most Common)

Ask only what you need to know about the user's side of the engagement. Skip target market and ICP questions — the target is already identified. Focus on understanding what problem your offering solves.

Your Business & Problem You Solve
- What does your company do? (elevator pitch)
- What problem does your offering solve? (be specific — not "improve efficiency" but the actual pain)
- What does life look like for a prospect before they work with you? (the pain state)
- What does life look like after? (the solved state)
- Why should they choose you over doing nothing or using an alternative?
Engagement Context
- What is the goal? (advisory engagement, product sale, partnership, tender response)
- What triggered this interest in [Company]? (event, referral, news, cold opportunity)
- Have you engaged with this company before? What happened?
- Do you have any existing relationship or warm introduction path?
- Is there an upcoming event, deadline, or trigger driving urgency?
Qualification Criteria
- What signals would tell you this company has the problem you solve?
- What would make this company a bad fit? (disqualifying criteria)
- What level of authority do you typically need to engage? (C-suite, director, manager)
Output Preferences
- Any specific focus areas? (e.g., focus on their technology landscape, or on decision-makers)
- Do you need outreach templates included?
- Will this feed into a BEAM engagement? (affects qualification depth)

Then proceed directly to Phase 3: Deep-Dive Intelligence for that specific company.

When No Target Company Is Provided (Pipeline Mode)

If the user asks for broad prospect identification (e.g., "find companies in mining for us"), conduct the full discovery:

Your Business
- What does your company do? (elevator pitch)
- What products or services are you looking to sell or partner on?
- What is your unique value proposition — why should a prospect choose you over alternatives?
- What is your typical deal size and sales cycle length?
Target Market
- What industry or vertical are you targeting? (e.g., mining, utilities, transport, government)
- What company size are you targeting? (revenue range, employee count, geographic scope)
- Are there specific geographies or regions you are focused on?
- Are there any companies you already have in mind?
Ideal Customer Profile (ICP)
- What business problems does your solution solve?
- What does the ideal buyer look like? (role, department, seniority)
- What technologies, systems, or processes does a good-fit company typically have in place?
- Are there any disqualifying criteria? (too small, wrong industry, existing competitor contract)
Engagement Goals
- What is the goal of this research? (cold outreach, conference preparation, partnership exploration, tender response)
- Do you have an upcoming event, conference, or deadline driving this?
- What does a successful engagement look like? (meeting booked, proposal sent, partnership signed)
- How many prospects do you need?
Existing Intelligence
- Do you have any existing prospect lists, CRM data, or past proposals to build on?
- Have you engaged with any of these companies before? What happened?
- Are there any warm introductions or mutual connections you know of?
Output Preferences
- What format do you need? (HTML report, Markdown, CSV prospect list, outreach email drafts)
- How detailed should each prospect profile be?
- Do you need outreach templates customised per prospect?

Research Methodology

Phase 1: Define the Problem Domain

Before researching the company, establish clarity on what you're looking for:

Articulate the problem hypothesis — what specific problem do you believe this company has?
Define problem signals — what evidence would indicate they have this problem? (job postings, news, technology gaps, regulatory pressures)
Define disqualifying signals — what would tell you they don't have this problem or are a poor fit?
Set research scope — depth of research, time allocated

Phase 2: Problem Signal Identification

Research with a problem-first lens — actively seek evidence that validates or invalidates your problem hypothesis:

Pain signals in news — incidents, complaints, regulatory issues, operational challenges
Job postings — roles that signal they're trying to solve problems you address
Technology gaps — systems that create problems your offering addresses
Strategic priorities — initiatives where your problem domain is relevant
Industry benchmarks — are they underperforming peers in your problem area?
Executive commentary — what are leaders saying about challenges?

Be honest: If you don't find evidence of the problem, document that. Qualifying out is valuable.

Phase 3: Deep-Dive Intelligence

For each shortlisted prospect, compile with a problem-centric focus:

Category	Details to Capture	BEAM Relevance
Company Overview	Name, HQ, revenue, employees, industry, sub-vertical	Context for fit assessment
Business Model	What they do, who they serve, how they make money	Understanding their world
Problem Evidence	Specific evidence that they have the problem you solve — not assumptions	Gate 1: Problem domain
Pain Quantification	What is this problem costing them? (dollars, time, risk, reputation)	BEAM Stage 2 preparation
Technology Landscape	Current systems and their limitations — problems created by their tech stack	Problem evidence
Buying Committee	Stakeholders mapped to BEAM buying roles (see below)	Gate 2: Access to authority
Access Path	How can you reach authority? Warm intros, events, content, cold	Gate 2 support
Willingness Signals	Evidence they're open to conversation — responding to content, attending events, issuing RFPs	Gate 3: Willingness
Competitive Landscape	Current vendors — are they happy or is there vendor fatigue/dissatisfaction?	Problem/timing signal
Qualification Assessment	Honest evaluation: is this a real opportunity or are we forcing it?	BEAM integrity

Decision-Maker Identification via LinkedIn (CRITICAL)

You MUST use web search with site:linkedin.com/in to find and identify key decision-makers. This is a core part of every dossier.

Search strategy — run these searches for each target company:

C-Suite and senior leadership:

site:linkedin.com/in "[Company Name]" CEO OR "Managing Director" OR "General Manager"

Operational / technical leaders (adapt titles to the offering):

site:linkedin.com/in "[Company Name]" "Head of" OR "Director" OR "VP" operations OR technology OR engineering OR "asset management"

Financial decision-makers:

site:linkedin.com/in "[Company Name]" CFO OR "Chief Financial" OR "Finance Director" OR "Head of Finance"

Procurement / buying roles:

site:linkedin.com/in "[Company Name]" procurement OR purchasing OR "vendor management" OR "strategic sourcing"

Domain-specific roles (tailor to what you're selling):

site:linkedin.com/in "[Company Name]" "digital transformation" OR "innovation" OR "data" OR "analytics"

For each decision-maker found, capture:

Field	What to Record
Full Name	As shown on their LinkedIn profile
Title	Current job title
Department	Inferred from title (e.g., Operations, IT, Finance, Strategy)
LinkedIn URL	Direct link to their profile (e.g., `https://linkedin.com/in/firstname-lastname`)
Relevance	Why this person matters for the engagement (budget authority, technical evaluator, operational sponsor, etc.)
Background Notes	Career history, prior companies, education, shared connections — anything useful for personalisation
Engagement Angle	How to approach this person specifically (what to reference, what pain point to lead with)

Map the BEAM buying committee — identify at minimum:

BEAM Role	Description	Typical Titles	What They Care About
Economic Buyer	Controls the budget; makes the final financial decision	CEO, CFO, VP, GM	ROI, risk, strategic fit
Technical Evaluator	Assesses solution fit and technical feasibility	CTO, Head of IT, Engineering Director	Will it work? Integration?
Champion / Sponsor	Internally advocates for your solution; feels the pain most	Director, Senior Manager, Program Lead	Solving their problem
Gatekeeper	Controls access to decision-makers and process	EA, Procurement Manager, PMO	Process, compliance, vendor management

CRITICAL: For BEAM Stage 1, you need at least one named individual you can access — not just "someone in IT". The skill will challenge vague answers.

Phase 4: BEAM Qualification Assessment

Before producing engagement strategy, honestly assess BEAM Stage 1 gate readiness:

Gate 1: Problem Domain Identified

Assessment	Criteria
STRONG	Multiple specific problems evidenced with sources; problems align directly to your offering
MODERATE	Some problem signals found but not highly specific; requires validation in discovery
WEAK	Assumed problems based on industry norms; no company-specific evidence
NONE	No evidence of the problem you solve; consider qualifying out

Gate 2: Access to Authority

Assessment	Criteria
STRONG	Named individual identified; existing relationship or warm introduction path
MODERATE	Named individual identified; no relationship but clear outreach path
WEAK	Roles identified but no specific names; requires LinkedIn research
NONE	No path to authority visible; consider qualifying out

Gate 3: Willingness to Diagnose

Assessment	Criteria
STRONG	Active signals: issued RFP, attending your events, responded to content
MODERATE	Passive signals: in relevant industry groups, hiring in your area
WEAK	No signals but no negative indicators either
NONE	Negative signals: recently bought competitor, stated no interest

Qualification Verdict

Based on the gate assessment, issue a verdict:

Verdict	Criteria	Recommended Action
QUALIFIED	2+ gates STRONG, none NONE	Proceed to engagement strategy
EXPLORATORY	Mixed signals; worth pursuing with lower investment	Light-touch outreach; validate problem first
NURTURE	Weak signals; not ready now	Add to content nurture; revisit in 6 months
DISQUALIFIED	Any gate NONE; fundamental misalignment	Document reasons; do not pursue

Phase 5: Engagement Strategy (If Qualified)

For qualified prospects, produce:

BEAM Gate Readiness — assessment of each Stage 1 gate with supporting evidence
Fit Score (1–5) based on ICP alignment and problem evidence strength
Timing Score (1–5) based on buying signals and urgency
Qualification Strength — STRONG, MODERATE, or EXPLORATORY
First Contact Goal — earn the right to a diagnostic conversation (not sell!)
Problem-Led Talking Points — 3–5 points leading with their problem, not your solution
Discovery Questions — SPIN-aligned questions to validate problem in first conversation
Outreach Sequence — multi-touch plan focused on earning the right to a conversation

Output Formats

1. Executive Opportunity Summary

Always lead with this summary before presenting the full dossier. It gives the user an immediate read on whether the prospect qualifies and at what confidence level.

# [Company Name] — Engagement Opportunity

## BEAM Qualification Readiness

| Gate | Status | Evidence Summary |
|------|--------|-----------------|
| **Problem Domain** | [STRONG/MODERATE/WEAK/NONE] | [1-line evidence summary] |
| **Access to Authority** | [STRONG/MODERATE/WEAK/NONE] | [Named contact or access path] |
| **Willingness to Diagnose** | [STRONG/MODERATE/WEAK/NONE] | [Signal or barrier] |

**Qualification Verdict: [QUALIFIED / EXPLORATORY / NURTURE / DISQUALIFIED]**

---

## Key Findings

🎯 **Problem We Can Solve**
- [Specific problem identified — evidenced, not assumed]
- [Impact of this problem on their business]
- [Why your offering addresses this problem specifically]

⏰ **Timing Signals**
- [Evidence of urgency — deadline, event, initiative]
- [Window of opportunity]
- [Risk if timing is missed]

🔑 **Buying Committee**
- **Economic Buyer**: [Name] ([Title]) — [access status]
- **Champion**: [Name] ([Title]) — [relationship status]
- **Technical Evaluator**: [Name] ([Title]) — [relevance]
- **Gatekeeper**: [Name/role] — [known or unknown]

💡 **Access Path**
- [Best route to a conversation — warm intro, event, content, cold]
- [Relationship leverage points]
- [Barriers to access and mitigation]

📊 **Qualification Confidence**
- [Fit Score]: [X/5] — [1-line rationale]
- [Timing Score]: [X/5] — [1-line rationale]
- [Overall]: [HIGH/MEDIUM/LOW] confidence this warrants pursuit

---

**First Contact Goal**: Earn the right to a diagnostic conversation — validate problem hypothesis, not pitch solution.

**If DISQUALIFIED**: [Reason] — document and move on. Qualifying out is a success.

2. Prospect Intelligence Dossier (Per Company)

The full detailed output for deep research on a single prospect.

# [Company Name] — Engagement Dossier

## Company Snapshot

| Field | Detail |
|-------|--------|
| **Company** | [Name] |
| **Website** | [URL] |
| **Headquarters** | [City, State/Country] |
| **Industry** | [Primary vertical] |
| **Revenue** | [Annual revenue or range] |
| **Employees** | [Headcount or range] |
| **Ownership** | [Public/Private/Government] |

## Business Overview

[2–3 paragraph summary of what the company does, who they serve, and their market position.]

## Strategic Priorities

- [Priority 1 — with source/evidence]
- [Priority 2 — with source/evidence]
- [Priority 3 — with source/evidence]

## Pain Points & Challenges

- [Pain point 1 — how it relates to your offering]
- [Pain point 2 — how it relates to your offering]
- [Pain point 3 — how it relates to your offering]

## Technology Landscape

| System | Vendor/Product | Notes |
|--------|---------------|-------|
| ERP | [e.g., SAP S/4HANA] | [Implementation date, known issues] |
| EAM | [e.g., IBM Maximo] | [Version, satisfaction level] |
| CRM | [e.g., Salesforce] | [Usage context] |
| Other | [Relevant systems] | [Notes] |

## Decision-Makers

*Found via `site:linkedin.com/in` web search*

| Name | Title | Department | Buying Role | Relevance | LinkedIn |
|------|-------|------------|-------------|-----------|----------|
| [Name] | [Title] | [Dept] | Economic Buyer | [Why they matter] | [linkedin.com/in/...] |
| [Name] | [Title] | [Dept] | Technical Evaluator | [Why they matter] | [linkedin.com/in/...] |
| [Name] | [Title] | [Dept] | Champion / Sponsor | [Why they matter] | [linkedin.com/in/...] |
| [Name] | [Title] | [Dept] | Gatekeeper | [Why they matter] | [linkedin.com/in/...] |

## Buying Signals

- [Signal 1 — date, source, interpretation]
- [Signal 2 — date, source, interpretation]

## Competitive Intelligence

| Competitor | Product/Service | Relationship Status | Vulnerability |
|-----------|----------------|--------------------:|--------------|
| [Vendor] | [Product] | [Active/Expiring/Dissatisfied] | [Why you could displace] |

## Engagement Strategy

### Fit Score: [X/5] | Timing Score: [X/5] | Priority: [HIGH/MEDIUM/LOW]

### Recommended Approach
[Specific approach — e.g., "Warm introduction via [mutual contact] targeting [decision-maker] at [upcoming event]"]

### Personalised Talking Points
1. [Point linking your value prop to their specific situation]
2. [Point referencing their recent initiative/announcement]
3. [Point addressing a known pain point]
4. [Point differentiating from their current vendor]

### Potential Objections & Responses
| Objection | Response |
|-----------|----------|
| "[Expected objection]" | "[How to counter]" |

### Outreach Sequence
| Step | Channel | Timing | Action |
|------|---------|--------|--------|
| 1 | Email | Day 1 | [Initial outreach — personalised cold email] |
| 2 | LinkedIn | Day 3 | [Connection request with note] |
| 3 | Email | Day 7 | [Follow-up with value-add content] |
| 4 | Phone | Day 10 | [Call script — reference email] |
| 5 | Email | Day 14 | [Break-up email or pivot to different contact] |

---

*Research compiled: [Date]*
*Sources: [List key sources]*

3. Prospect Pipeline (Multi-Company)

For batch research across multiple prospects.

# B2B Prospect Pipeline — [Target Vertical / Campaign Name]

## Research Parameters

- **Your Company**: [Name]
- **Offering**: [Product/Service]
- **Target Vertical**: [Industry]
- **Target Geography**: [Region]
- **ICP Criteria**: [Summary]
- **Date**: [Research date]

## Pipeline Summary

| # | Company | Industry | Revenue | Fit | Timing | Priority | Recommended Approach |
|---|---------|----------|---------|-----|--------|----------|---------------------|
| 1 | [Name] | [Vertical] | [Rev] | [1-5] | [1-5] | HIGH | [Approach] |
| 2 | [Name] | [Vertical] | [Rev] | [1-5] | [1-5] | HIGH | [Approach] |
| 3 | [Name] | [Vertical] | [Rev] | [1-5] | [1-5] | MEDIUM | [Approach] |

## Tier 1 Prospects (High Priority)

### [Company 1]
[Brief paragraph — why they are Tier 1, key trigger, recommended next step]

**Key Contact**: [Name, Title]
**Trigger**: [What makes this timely]
**Next Step**: [Specific action]

### [Company 2]
...

## Tier 2 Prospects (Medium Priority)

### [Company 3]
...

## Tier 3 Prospects (Monitor)

### [Company N]
...

## Recommended Actions (Next 30 Days)

1. [Action 1 — specific, measurable, with owner and deadline]
2. [Action 2]
3. [Action 3]

---

*Pipeline compiled: [Date]*

4. Outreach Email Templates

Personalised email sequences for each prospect.

# Outreach Templates — [Company Name]

## Email 1: Initial Outreach

**Subject**: [Personalised subject line referencing their situation]
**To**: [Decision-maker name and title]

---

Hi [First Name],

[Opening line referencing a specific trigger — their recent announcement, a shared connection, or industry event.]

[1–2 sentences connecting their challenge to your value proposition. Be specific to their situation, not generic.]

[Social proof — a brief reference to a similar company you've helped, with a measurable outcome.]

[Clear, low-commitment call to action — e.g., "Would a 15-minute call next week make sense to explore whether this could work for [Company]?"]

Best regards,
[Your Name]
[Your Title] | [Your Company]
[Phone] | [Email]

---

## Email 2: Follow-Up (Day 7)

**Subject**: Re: [Original subject] — [Value-add hook]

---

Hi [First Name],

[Reference previous email briefly — don't guilt-trip.]

[Provide genuine value — share a relevant insight, article, case study, or data point that's useful to them regardless of whether they buy.]

[Reiterate the CTA with a slight variation.]

Best regards,
[Your Name]

---

## Email 3: Break-Up (Day 14)

**Subject**: Should I close your file?

---

Hi [First Name],

[Acknowledge they're busy. No pressure.]

[One final value statement — "If [specific outcome] is still a priority for [Company], I'd welcome the chance to share how [your company] helped [similar company] achieve [result]."]

[Permission-based close — "If the timing isn't right, no worries at all. I'll keep an eye out for anything relevant to share."]

Best regards,
[Your Name]

5. Conference / Event Prep Brief

For preparing engagement strategies around a specific event.

# Event Engagement Brief — [Event Name]

## Event Details

| Field | Detail |
|-------|--------|
| **Event** | [Name] |
| **Date** | [Date range] |
| **Location** | [Venue, City] |
| **Website** | [URL] |
| **Our Presence** | [Exhibiting/Attending/Speaking/Sponsoring] |

## Target Attendees

| # | Company | Contact | Title | Booth/Session | Engagement Plan |
|---|---------|---------|-------|---------------|----------------|
| 1 | [Company] | [Name] | [Title] | [Location] | [Approach] |

## Pre-Event Outreach

[Email templates to send before the event requesting meetings.]

## On-Site Talking Points

### For [Company 1]
- [Point 1]
- [Point 2]
- [Point 3]

## Post-Event Follow-Up

[Follow-up email template referencing the conversation at the event.]

6. HTML Report (Standalone Web Document)

The preferred delivery format. A polished, branded HTML report with SAS-AM styling, light/dark mode toggle, sidebar navigation, and interactive elements. Opens directly in any browser.

File Structure

report-folder/
├── report.html          # Main report file
├── styles.css           # Report theme styles (from references/report-styles.css)
└── assets/              # Logos and images
    ├── sas-logo-light.png
    └── sas-logo-dark.png

How to Build

Create a new directory for the report
Copy report-styles.css from references as styles.css
Copy report-template.html from references as the scaffold
Replace all {{PLACEHOLDER}} values with research findings
Add or remove sections, table rows, cards, and timeline steps as needed
Place logo assets in assets/

Key Features

Light/Dark Mode: Toggle via header button, saved to localStorage
Sidebar Navigation: Sticky nav with scroll-tracking highlights (IntersectionObserver)
Executive Summary Cards: Grid of finding cards with icons (🎯 ⏰ 💡 🔑 📊)
Score Badges: Colour-coded circular badges (1–5) for Fit and Timing scores
Priority Badges: HIGH (red), MEDIUM (amber), LOW (grey) pill badges
Outreach Timeline: Vertical timeline with day markers and channel badges
Summary Callout: Green-bordered highlight box for the executive conclusion
Responsive: Collapses sidebar on small screens, stacks cards on mobile
Print-Ready: Hides interactive elements, clean page layout

Design System

Uses the same SAS-AM brand system as the presentation skill:

Element	Light Mode	Dark Mode
Brand Blue	`#002244`	`#1a4d7a`
Brand Green	`#69BE28`	`#7AD33B`
Background	`#ffffff`	`#0a0f1a`
Text Primary	`#002244`	`#f4fbff`
Card Background	`#ffffff`	`#141d2b`

Typography: Source Sans Pro (300, 400, 600, 700 weights) from Google Fonts Icons: Font Awesome 6.5.1 from CDN

Scoring Rubric (CRITICAL — Pessimistic by Default)

Philosophy: Default to skepticism. A prospect must earn a high score through concrete evidence, not assumptions or optimism. If you're uncertain, score down. It's better to undersell a qualified opportunity than to overcommit to a dud.

Fit Score (1–5) — Assume Low Until Proven Otherwise

Score	Criteria	Evidence Required	Red Flags That Drop Score
5 — Perfect Fit	Has documented evidence of the exact problem you solve; matches ALL ICP criteria with proof	Direct quotes, specific incident reports, job postings explicitly describing the problem, regulatory filings showing the issue	Never assume 5 without multiple independent evidence sources
4 — Strong Fit	Has multiple signals of the problem; matches most ICP criteria	At least 2 independent evidence sources confirming problem domain	Single evidence source = max score 3
3 — Moderate Fit	Some evidence of potential problem; matches core ICP criteria	At least 1 concrete signal; plausible fit argument	Industry-level assumptions without company-specific evidence
2 — Weak Fit	Industry suggests potential fit but no company-specific evidence of the problem	ICP match on paper only	Assuming need based on what "companies like them" typically have
1 — Poor Fit	No evidence of problem; disqualifying criteria present	N/A — this is the absence of evidence	Evidence of recent competitor purchase; stated no need; misaligned industry

Default starting point: Score 2 until evidence moves it up. Optimism doesn't earn points.

Timing Score (1–5) — Assume Cold Until Proven Hot

Score	Criteria	Evidence Required	Red Flags That Drop Score
5 — Immediate	Active RFP mentioning your problem domain; documented budget allocation; leadership statement with deadline	Written RFP, board announcement with timeline, regulatory deadline	"They're probably looking" = max score 2
4 — Near-Term	Recent trigger event specifically related to your problem domain; decision expected within 6 months	Job posting for role that would use your solution; incident report; leadership change with relevant mandate	Trigger event unrelated to your problem domain = score 3 max
3 — Medium-Term	Strategic alignment visible; no confirmed timeline; may be 6–12 months	Strategic plan mentions your problem area; preliminary discussions reported	No timeline visibility = cap at 3 regardless of fit
2 — Long-Term	Potential future need based on industry trends; no company-specific signals	Industry reports suggesting eventual need	Assuming timing based on "everyone will need this eventually"
1 — No Signal	No buying signals detected; no trigger events; no urgency indicators	N/A — this is the absence of evidence	Evidence of recent purchase or "not a priority" statement

Default starting point: Score 2 until a specific, dated trigger event moves it up.

Qualification Rigour Tests

Before assigning Fit or Timing scores, answer these questions honestly:

Test	Question	If "No"...
Specificity	Is this evidence specific to THIS company, or just industry-level assumptions?	Cap score at 2
Recency	Is this evidence from the last 12 months?	Reduce score by 1
Independence	Do you have 2+ independent sources?	Cap Fit at 3
Problem Match	Does the evidence directly match YOUR problem domain, or just adjacent?	Cap Fit at 3
Timeline Clarity	Is there a specific date, deadline, or decision event?	Cap Timing at 3
Access Reality	Can you actually reach a decision-maker, or are you hoping?	Note barrier in assessment

Priority Matrix (Conservative)

Default to NURTURE unless evidence compels higher priority. Pursuing a weak opportunity costs more than missing a marginal one.

	Timing 4–5 (Hot)	Timing 3 (Warm)	Timing 1–2 (Cold)
Fit 5	HIGH — Pursue immediately; rare — verify twice	MEDIUM — Nurture actively; timing may shift	LOW — Monitor; great fit but no urgency
Fit 4	HIGH — Strong case; validate access	MEDIUM — Worth light pursuit	NURTURE — Build relationship
Fit 3	MEDIUM — Validate problem first	NURTURE — Stay in touch	SKIP — Not worth the effort
Fit 1–2	EXPLORATORY at best — validate fit first	SKIP	SKIP

Evidence Quality Tiers

Tier	Evidence Type	Score Multiplier
Tier 1 (Gold)	Direct quote from prospect; their written RFP; their public filing mentioning the problem	Full score valid
Tier 2 (Silver)	News article about them + the problem; job posting; investor report mentioning challenge	Score - 1 max unless corroborated
Tier 3 (Bronze)	Industry analyst report; peer company comparison; your assumption based on typical patterns	Cap at 3 regardless of logic
Tier 4 (Lead)	"Companies like them usually..." / "They must have..." / "I assume..."	Cap at 2; flag as unverified

Honest Assessment Output

For every Fit and Timing score, document:

Fit Score: X/5
- Evidence: [What specific evidence supports this?]
- Evidence Tier: [Gold/Silver/Bronze/Lead]
- Confidence Killer: [What would disprove this?]
- Honest Doubt: [What aren't you sure about?]

Timing Score: X/5
- Evidence: [What specific trigger/signal supports this?]
- Evidence Tier: [Gold/Silver/Bronze/Lead]
- Timeline Uncertainty: [What could delay or accelerate?]
- Honest Doubt: [What aren't you sure about?]

Overall Assessment: [QUALIFIED/EXPLORATORY/NURTURE/DISQUALIFIED]
Confidence Level: [HIGH/MEDIUM/LOW]
If Wrong: [What's the cost if this assessment is wrong?]

Research Sources

When conducting research, prioritise these source types:

Primary Sources (Most Reliable)

Company website (About, Careers, News/Blog, Investor Relations)
Annual reports and investor presentations
SEC filings (for public companies)
Government procurement portals and tender notices
Official press releases

LinkedIn Sources (Decision-Maker Identification — REQUIRED)

site:linkedin.com/in "[Company]" [title] — find specific people by company and role
site:linkedin.com/company "[Company]" — company page for employee count, industry, and overview
site:linkedin.com/in "[Company]" "joined" OR "started" — recent hires signalling investment areas
site:linkedin.com/posts "[Company]" [topic] — stakeholder posts revealing priorities and opinions

Secondary Sources (Cross-Reference)

Industry publications and trade journals
Conference programmes and speaker lists
Industry association member directories
News aggregators and business media

Signal Sources (Buying Intent)

Job postings (especially for roles in your solution area)
Technology review sites (G2, Gartner Peer Insights, Capterra)
Patent filings and R&D announcements
Regulatory filings and compliance reports
Social media activity (LinkedIn posts from key stakeholders)

Workflow

Path A: Target Company Provided (Single-Company Dossier)

Use this path when the user provides a specific company name (the most common invocation).

Step 1: Problem-First Discovery

Ask the focused discovery questions: what problem do you solve? what would evidence of that problem look like?
Confirm the target company name and establish a problem hypothesis
Clarify: what would make this company a bad fit?

Step 2: Problem Evidence Research

Search for evidence that validates the problem hypothesis:
- News articles mentioning challenges, incidents, or pain in your problem domain
- Job postings for roles that signal they're trying to solve this problem
- Technology limitations that create the problem you address
- Executive statements about challenges or priorities
- Industry benchmarks showing underperformance
Be honest: Document both confirming AND disconfirming evidence
If no problem evidence found, document that — qualifying out is valid

Step 3: Buying Committee Mapping

Search LinkedIn for decision-makers using site:linkedin.com/in "[Company]" [title] queries:
- Search for Economic Buyers (CEO, CFO, GM, VP)
- Search for Technical Evaluators (CTO, Head of Engineering, IT Director)
- Search for potential Champions (directors and managers in the problem domain)
- Search for Gatekeepers (procurement, EA, PMO)
For each person found: capture name, title, LinkedIn URL, BEAM buying role, and access assessment
Identify the most viable access path — who can you actually reach?

Step 4: BEAM Qualification Assessment

Assess each Stage 1 gate:
- Problem Domain: Is there specific, evidenced problem or just assumptions?
- Access to Authority: Do you have a named individual and a path to reach them?
- Willingness: What signals suggest they'd engage in a conversation?
Issue a qualification verdict: QUALIFIED, EXPLORATORY, NURTURE, or DISQUALIFIED
If DISQUALIFIED: Document the reason and stop — this is valuable intelligence, not failure

Step 5: Engagement Strategy (If Qualified)

Score fit and timing (1–5 each)
Develop problem-led talking points (lead with their pain, not your features)
Prepare SPIN discovery questions for the first conversation
Design outreach focused on earning the right to a diagnostic conversation
Anticipate barriers and prepare to address them

Step 6: Delivery

Build the branded HTML report from the template
Lead with BEAM Qualification Readiness — gate assessments and verdict
Include the Executive Opportunity Summary
Include full dossier, strategy, and outreach plan
First Contact Goal: Earn the right to a diagnostic conversation
Provide recommended next steps with dates
If DISQUALIFIED: Deliver a brief disqualification report explaining why — this is still valuable output

Path B: No Target Company (Pipeline Mode)

Use this path when the user asks for broad prospect identification across a market.

Step 1: Full Discovery

Conduct the complete 6-section discovery interview
Confirm ICP criteria, target vertical, geography, and scope

Step 2: Research Planning

Define the research scope (number of prospects, depth, geography)
Establish the scoring rubric (confirm or adjust default criteria)
Identify primary research sources for the target vertical

Step 3: Prospect Identification

Use web search to identify candidate companies
Apply ICP filters to create a shortlist
Validate each candidate against qualifying criteria

Step 4: Deep-Dive Research For each shortlisted prospect:

Compile the company snapshot
Map decision-makers and buying committee
Identify pain points and engagement triggers
Assess competitive landscape
Score fit and timing

Step 5: Strategy Development

Rank prospects into tiers (High / Medium / Monitor)
Develop personalised engagement strategies per prospect
Draft outreach templates and talking points
Create a 30-day action plan

Step 6: Delivery

Build the HTML report using the branded template (report-template.html + report-styles.css):
- Create a report directory
- Copy report-styles.css from references as styles.css
- Copy report-template.html from references and populate all placeholders with research findings
- Add/remove sections, table rows, and cards as needed
- Place logo assets in assets/
Review for accuracy and completeness
Highlight the top 3 recommended next steps
Deliver the final report (HTML by default, Markdown if requested)

Content Writing Guidelines

Dossier Tone

Professional and analytical — present findings as intelligence, not opinion
Evidence-based — cite sources and dates for all claims
Actionable — every insight should connect to a recommended action
Concise — use tables for structured data, paragraphs for analysis

Outreach Tone

Personalised — reference specific company details, never generic
Value-first — lead with insight, not a sales pitch
Respectful — acknowledge their time; no pressure tactics
Credible — include specific social proof with measurable outcomes
Australian English — use Australian spelling conventions (organisation, colour, behaviour, prioritise)

Formatting Standards

Default output is an HTML report using the branded report template (report-template.html + report-styles.css)
Fall back to Markdown if the user specifically requests it
Use tables for structured/comparative data
Use bullet points for lists of findings
Use headings (H2/H3) to organise sections
Include a date stamp on all deliverables
Link to sources where possible

Engagement Timeline Generation

The engagement timeline provides a chronological view of all activities related to a prospect. It complements the kanban board by showing the full history of engagement in an easy-to-scan format.

When to Generate Timeline

Generate or update the timeline HTML file:

After initial research dossier is complete
When major activities occur (outreach, meetings, documents, decisions)
At the end of each engagement session
When advancing to a new BEAM stage

Timeline File Structure

Store the timeline in the engagement directory:

.beam/
└── engagements/
    ├── {company}-timeline.html    # Visual timeline (HTML)
    ├── {company}-kanban.html      # Kanban board (HTML)
    └── {company}.json             # Engagement state (JSON)

Timeline Entry Requirements

Each timeline entry MUST include:

Field	Required	Description
`date`	Yes	Specific date in DD MMM YYYY format (e.g., "23 Feb 2026")
`type`	Yes	One of: outreach, meeting, document, evidence, decision, milestone, blocker
`stage`	Yes	BEAM stage number (1-6)
`title`	Yes	Specific action title — NOT generic (see specificity guidelines below)
`description`	Yes	1-2 sentences explaining what happened
`details`	Recommended	Structured key:value pairs with specifics (channel, recipient, file, etc.)
`outcome`	Recommended	success, pending, or blocked with description

Timeline Entry Specificity (CRITICAL)

BAD (too generic):

title: "Outreach to James Hicks"
description: "Sent message requesting catch-up"

GOOD (specific and actionable):

title: "LinkedIn message sent to James Hicks"
description: "Sent personalised message referencing AMPEAK2024 conference and Bega consolidation news, requesting coffee catch-up to discuss asset management challenges."
details:
  Channel: LinkedIn
  Recipient: James Hicks (Group Manager Asset Management)
  Message: "Hi James, Hope you're well! I've been following the Bega consolidation news..."
outcome: pending — Awaiting response

Timeline Template

Use references/engagement-timeline-template.html as the scaffold. Replace placeholders:

Placeholder	Value
`{{COMPANY_NAME}}`	Prospect company name
`{{CURRENT_STAGE}}`	BEAM stage number (1-6)
`{{WIN_PROBABILITY}}`	Current win probability percentage
`{{DAYS_ACTIVE}}`	Days since engagement started
`{{ACTIVITY_COUNT}}`	Total number of timeline entries
`{{GENERATED_DATE}}`	Current date
`{{TIMELINE_ENTRIES}}`	HTML for each timeline entry (see template comments)

Kanban Board Management

The kanban board visualises current activities and BEAM gate progress. It should be updated immediately when activities are proposed, started, completed, or blocked.

Kanban Update Triggers

Update the kanban board when:

Trigger	Action
Research completed	Add card with status "done"
Outreach sent	Add card with status "doing", awaiting response
Meeting scheduled	Add card with status "todo", include date
Meeting held	Move card to "done", add outcome notes
Document created	Add card with status "done", reference file
Evidence captured	Add card with status "done", link to gate
Blocker identified	Add card with status "blocked", describe issue
Gate criteria met	Update gate in JSON, add evidence card
Stage transition	Move cards to archive, update stage status

Kanban Card Specificity (CRITICAL)

Every kanban card MUST be specific enough that someone reading it understands exactly what happened without needing to investigate further.

BAD EXAMPLES:

Card	Problem
"Research complete"	What research? What was found?
"Outreach to James"	What channel? What was said? When?
"Meeting scheduled"	With whom? When? What's the agenda?
"Document created"	What document? Where is it?

GOOD EXAMPLES:

Card	Why It Works
"B2B dossier ingested — Fit: 5/5, Timing: 5/5, 21 manufacturing sites identified"	Specific outcome, quantified
"LinkedIn message sent to James Hicks 23 Feb — referenced AMPEAK2024, requested coffee"	Channel, date, content summary
"Discovery call scheduled: James Hicks, 28 Feb 2pm AEDT, Teams"	All logistics clear
"Pilot proposal drafted: bega-pilot-proposal.html — $35K, 5 weeks, 2-3 sites"	File reference, commercial terms

Card Notes Requirements

The notes field on each kanban card should include:

What was done — specific action taken
Evidence captured — any quotes, data, or signals gathered
Outcome/status — result of the action
Next step dependency — what's waiting on this card

Example:

{
  "id": "1-004",
  "type": "meeting",
  "title": "LinkedIn message sent to James Hicks 23 Feb",
  "status": "doing",
  "notes": "Sent personalised message via LinkedIn referencing AMPEAK2024 and Bega consolidation. Message: 'Hi James, Hope you're well! I've been following the Bega consolidation news...' Awaiting response. Follow up Day 3 (26 Feb) if no reply. Gate: access_to_authority depends on positive response.",
  "gate_ref": "access_to_authority",
  "created_at": "2026-02-23"
}

Kanban State Consistency

The kanban board and engagement JSON must stay in sync:

activity_log[] in JSON should match kanban cards
timeline.milestones[] should reflect completed kanban cards
stages.{stage}.evidence[] should reference evidence-type cards
kanban.{stage}.cards[] should have correct status reflecting JSON state

Activity Update Protocol

When ANY of the following activities occur, you MUST update both the kanban board and timeline:

Immediate Update Activities

Activity	Kanban Update	Timeline Update
Outreach sent (email, LinkedIn, call)	Add card: type=meeting, status=doing	Add entry: type=outreach
Response received	Update card status to done/blocked	Add entry: type=evidence or blocker
Meeting scheduled	Add card: type=meeting, status=todo	Add entry: type=meeting (scheduled)
Meeting held	Update card status to done	Add entry: type=meeting (held)
Document created	Add card: type=document, status=done	Add entry: type=document
Evidence captured	Add card: type=evidence, status=done	Add entry: type=evidence
Gate criteria met	Update gate in JSON, add evidence card	Add entry: type=milestone
Stage transition	Archive previous stage cards	Add stage-transition block
Blocker identified	Add card: type=blocker, status=blocked	Add entry: type=blocker
Decision made	Add card: type=evidence, status=done	Add entry: type=decision

Update Process

Identify activity type from the list above
Capture specifics — date, channel, participants, content summary, outcome
Update engagement JSON — add to activity_log[], update relevant stage, add to timeline.milestones[]
Update kanban HTML — add/update card in appropriate stage column
Update timeline HTML — add entry with full details
Save all files — JSON, kanban HTML, timeline HTML

Session Summary Updates

At the END of every session, update the session log:

# Session Log — {{COMPANY}} — {{DATE}}

## What We Covered
- [List of activities completed this session]

## Evidence Collected
| ID | Type | Description | Date | Supports |
|----|------|-------------|------|----------|
| ev-XXX | ... | ... | ... | ... |

## Gate Progress
| Gate | Verdict | Evidence |
|------|---------|----------|
| ... | ... | ... |

## Next Steps
1. [Specific next action with date]
2. ...

## Current State
- Stage: X
- Win Probability: X%
- Gates Met: X/X

Context Orchestration Protocol (CRITICAL)

Long-running engagements span multiple sessions. Use this protocol to manage context windows by saving work to files and resuming cleanly.

The Problem

Context windows are limited
Sales engagements can span weeks or months
Without orchestration, context is lost between sessions
Duplicate work occurs when previous context isn't loaded

The Solution: Save → Clear → Resume Pattern

┌─────────────────────────────────────────────────────────────┐
│  SESSION START                                               │
│  ┌─────────────────────────────────────────────────────┐    │
│  │ Load Context from Files                               │    │
│  │ • Read {company}.json                                 │    │
│  │ • Read latest session log                             │    │
│  │ • Review kanban board state                           │    │
│  └─────────────────────────────────────────────────────┘    │
│                           ▼                                  │
│  ┌─────────────────────────────────────────────────────┐    │
│  │ Work in Session                                       │    │
│  │ • Perform activities                                  │    │
│  │ • Capture evidence                                    │    │
│  │ • Update state incrementally                          │    │
│  └─────────────────────────────────────────────────────┘    │
│                           ▼                                  │
│  ┌─────────────────────────────────────────────────────┐    │
│  │ Save Context to Files (CRITICAL)                      │    │
│  │ • Update {company}.json with all changes              │    │
│  │ • Update kanban HTML with new cards                   │    │
│  │ • Update timeline HTML with new entries               │    │
│  │ • Write session log with summary                      │    │
│  └─────────────────────────────────────────────────────┘    │
│  SESSION END                                                 │
└─────────────────────────────────────────────────────────────┘

Context Files Structure

{research_directory}/
├── .beam/
│   ├── config.json                     # Seller context (persists across all engagements)
│   ├── engagements/
│   │   ├── {company}.json              # MASTER STATE — all engagement data
│   │   ├── {company}-kanban.html       # Visual kanban board
│   │   └── {company}-timeline.html     # Visual timeline
│   └── sessions/
│       ├── {company}-{date}.md         # Session log (one per session)
│       └── ...
├── {company}-dossier.html              # Initial research dossier
├── {company}-proposal.html             # Proposal document (when created)
└── styles.css                          # Shared styles

Session Start Protocol

When resuming an engagement, ALWAYS:

Read the master JSON file — {company}.json contains full state
Read the latest session log — understand where we left off
Review current kanban state — what's in progress, blocked, or completed
Summarise context for user — "Last session we... Next steps were..."

Example prompt to user:

📋 **Resuming Bega Group engagement**

**Last session (23 Feb 2026):**
- LinkedIn message sent to James Hicks
- Pilot proposal drafted ($35K, 5 weeks)
- Conversation prep materials created

**Current state:**
- Stage: 1 — Qualify (1/3 gates passed)
- Win probability: 10%
- Pending: Awaiting James Hicks response

**Next steps from last session:**
1. Wait for James response
2. Follow up Day 3 (26 Feb) if no reply
3. Send value-add content Day 7

**What would you like to do?**

Session End Protocol (CRITICAL)

At the END of every session, ALWAYS:

Update master JSON — capture ALL changes to engagement state
Update kanban HTML — reflect current card states
Update timeline HTML — add all activities from this session
Write session log — summarise what was done, evidence collected, next steps
Confirm save — tell the user files have been updated

Example:

✅ **Session saved**

Updated files:
- `.beam/engagements/bega-group.json` — engagement state
- `.beam/engagements/bega-group-kanban.html` — 2 cards updated
- `.beam/engagements/bega-group-timeline.html` — 3 entries added
- `.beam/sessions/bega-group-2026-02-26.md` — session log

**Pick up next session by reading the session log.**

Activity Checkpoints

Some activities are natural checkpoints where you should save context:

Activity	Save?	Rationale
Research dossier complete	Yes	Major deliverable, captures all discovery work
Outreach sent	Yes	Activity logged, waiting for external response
Meeting completed	Yes	Evidence captured, next steps defined
Stage transition	Yes	Major milestone, previous stage archived
Document created	Yes	Asset saved to file system
Proposal delivered	Yes	Critical commercial milestone
User requests pause	Yes	Explicit checkpoint requested

Loading Previous Context

When loading context, prioritise files in this order:

{company}.json — master state (most authoritative)
Latest session log — recent activity summary
Kanban board — visual state verification
Timeline — chronological activity history

Minimal Context for Resume

To resume an engagement with minimal context load, read:

{
  "current_stage": 1,
  "win_probability": 10,
  "last_session": "2026-02-23",
  "gates_met": ["problem_domain_identified"],
  "gates_pending": ["access_to_authority", "willing_to_diagnose"],
  "last_activity": "LinkedIn outreach to James Hicks",
  "next_steps": ["Await response", "Follow up Day 3"]
}

This summary can be extracted from the master JSON and presented to the user without loading full context.

Checklist

Research Delivery Checklist

Before delivering initial research, verify:

Kanban Board Checklist

When updating the kanban board, verify:

All card titles are specific (include date, channel, recipient, content summary)
Card notes explain what was done, evidence captured, and outcome
Card status is accurate (todo, doing, done, blocked)
Gate references are linked for evidence-type cards
Stage progress bar reflects current gate criteria state
Activity log in JSON matches kanban cards
No generic card titles like "Outreach sent" or "Meeting held"

Timeline Checklist

When generating/updating timeline, verify:

All entries have specific dates (DD MMM YYYY format)
Entry types are correctly categorised (outreach, meeting, document, evidence, milestone, blocker)
Entry titles are specific (not generic)
Details section includes channel, recipients, content summary
Outcomes are marked (success, pending, blocked)
Stage transitions are highlighted as separate blocks
Filters work correctly
Entry count matches actual activities

Session End Checklist (CRITICAL)

Before ending ANY session, verify:

Session Start Checklist

When resuming an engagement, verify:

Master JSON file read and state understood
Latest session log read for context
User briefed on last session summary
Current state presented (stage, win probability, gates, pending activities)
Next steps from last session reviewed
User asked what they want to do this session