Available Context & Tools

@_platform-references/org-variables.md @_platform-references/capabilities.md

Lead Research

Goal

Research a lead or company using web search to gather actionable intelligence for sales outreach. The output should give a rep everything they need to write a personalized, relevant first touch -- and enough context to hold an informed conversation if the prospect responds.

Good lead research is not a data dump. It is a curated intelligence brief that answers three questions: Who is this person? What does their company do? And what should I say to them that will actually get a response?

Required Capabilities

• Web Search: To search the web for lead and company information (routed to Gemini with Google Search grounding)

Inputs

• lead_name : Name of the person to research (if available)

• company_name : Name of the company to research (if available)

• email : Email address of the lead (if available, useful for finding LinkedIn profiles and confirming identity)

• organization_id : Current organization context

Research Methodology

Lead research follows a three-phase approach: cast a wide net, go deep on what matters, then synthesize into actionable intelligence. See references/research-playbook.md for comprehensive search query templates, LinkedIn analysis techniques, time-boxed protocols (5/15/30 minute), and hiring signal interpretation.

Phase 1: Discovery (Run Searches in Parallel)

Launch 5-7 searches simultaneously to maximize coverage and minimize latency. The goal is breadth -- you want to discover all available data sources before committing to deep dives.

If you have a person name + company:

• "[Person Name]" "[Company Name]" LinkedIn -- find their professional profile

• "[Person Name]" "[Company Name]" OR [role keywords] -- find mentions, quotes, content

• "[Company Name]" company about -- official company information

• "[Company Name]" news OR announcement OR funding -- recent activity

• "[Company Name]" careers OR hiring OR "open roles" -- hiring signals and tech stack clues

• "[Person Name]" podcast OR speaker OR interview OR article -- thought leadership and content

• "[Company Name]" review OR G2 OR Capterra -- market reputation

If you have only a person name:

• "[Person Name]" LinkedIn [any known context] -- find the right profile

• "[Person Name]" [industry or city or any disambiguating info] -- narrow down identity

• If email domain is available, use domain to identify company, then expand searches

If you have only a company name:

• "[Company Name]" company about -- official info

• "[Company Name]" founders OR CEO OR leadership -- identify key people

• "[Company Name]" funding OR investors OR Crunchbase -- financial context

• "[Company Name]" news recent -- recent developments

• site:[company-domain.com] -- what they publish about themselves

If you have only an email:

• Extract the domain. If it's a corporate domain (not gmail/yahoo/outlook), search for the company.

• Search for the email address directly -- it may appear in press releases, conference bios, or public directories.

• Once you have a company, expand to the full search pattern above.

Phase 2: Deep Dive (Fetch Promising URLs)

Based on Phase 1 results, fetch and extract from the most valuable pages. Prioritize in this order:

• LinkedIn profile (highest value for person research) -- current role, tenure, employment history, education, activity, shared connections

• Company About/Team page -- official description, leadership, product lines, mission

• Company Pricing/Product page -- what they sell, pricing model, target market

• Crunchbase/PitchBook profile -- funding history, investors, valuation, growth metrics

• Recent blog posts or press releases (last 6 months) -- strategic direction, product launches, partnerships

• Job postings (current openings) -- tech stack, growth areas, team structure, priorities

• Conference bios or speaker pages -- speaking topics, areas of expertise, professional interests

• Review sites (G2, Capterra, Trustpilot) -- market reputation, strengths, weaknesses

Time budget guidance: Spend roughly 60% of research effort on the person, 40% on the company. The rep needs to connect with a human, not a logo. The company context exists to make the person conversation better.

Phase 3: Synthesis (Merge and Validate)

Compile all findings into the structured output. During synthesis:

• Cross-reference identity. Confirm that the person you found is actually the right person at the right company. Name collisions are common. Verify by cross-checking title, company, location, and timeline.

• Resolve conflicting data. If LinkedIn says "Senior Director" but the company website says "VP," note both and use the most recently updated source. The person's own LinkedIn is usually the freshest.

• Assess data freshness. Note when each piece of information was last updated. LinkedIn profiles updated in the last 3 months are high confidence. Company websites may not reflect recent changes.

• Identify gaps. Explicitly note what you could NOT find. "Phone number: not found via web search" is more useful than silently omitting the field.

• Extract connection points. The most valuable output of research is not data -- it's connection points. Things the rep can reference in outreach that show they did their homework: a recent post, a career move, a company announcement, a shared interest.

LinkedIn Intelligence Extraction

LinkedIn is the single most valuable source for lead research. Here is what to extract and why each matters:

Profile Basics

• Current title + company: Confirms identity and role

• Headline: Often more revealing than title -- people write their own headlines to signal what they care about

• Location: Time zone context for scheduling; local references in outreach

Career Trajectory

• Tenure at current company: <6 months = new in role (high openness to new tools). 2+ years = established (harder to change but knows the org well).

• Previous companies: If they came from a company in your customer base, mention it. If they came from a competitor's customer base, they may have familiarity with the category.

• Career pattern: Rapid promotions suggest a high performer. Lateral moves suggest breadth. Long tenures suggest loyalty. Frequent moves suggest restlessness. Each tells you something about how to approach them.

Content and Activity

• Recent posts: What topics do they post about? These are their professional interests and priorities.

• Articles published: Indicates thought leadership and areas of deep expertise.

• Comments and engagement: Even if they don't post, what do they react to? This reveals interests.

• Recommendations given/received: Reveals professional relationships and what people value about them.

Network Signals

• Shared connections: If anyone at ${company_name} is connected to this person, flag it immediately. Warm introductions are gold.

• Group memberships: Professional groups reveal community affiliations and interests.

• Followed companies/influencers: Reveals what they pay attention to in the market.

Company Research Shortcuts

When time is limited, here is where to find the best data fastest, in priority order:

• Company website /about page: 60 seconds for company description, founding year, leadership, mission

• LinkedIn company page: 30 seconds for employee count, growth trend, headquarters, industry

• Crunchbase: 30 seconds for funding, investors, key people, employee count history

• Job postings (careers page or LinkedIn jobs): 2 minutes for tech stack, growth areas, team priorities

• Google News search: 1 minute for recent announcements, funding, product launches

• G2/Capterra page: 1 minute for market category, competitor set, customer sentiment

Total: you can build a solid company profile in under 5 minutes if you know where to look.

News Analysis: Signal vs Noise

Not all company news is relevant for sales outreach. Here is how to filter:

High-Signal News (always include)

• Funding announcement: New capital = new budget, new hiring, new tool purchases. The 90 days after a funding round is the best window for vendor outreach.

• New product launch: Indicates strategic direction and where they're investing. If ${company_name}'s products (from Organization Context) enable their new product, that's a direct talking point.

• Leadership change: New CTO/VP Engineering = new tool stack review. New CEO = strategic pivot. New VP Sales = new process evaluation.

• Partnership/acquisition: Reveals strategic priorities and potential integration needs.

• Expansion (new office, new market, international growth): Signals budget, growth, and potential scaling challenges that ${company_name}'s solutions can address.

Medium-Signal News (include if relevant to ${company_name}'s products)

• Industry awards or rankings: Nice for flattery in outreach ("Congrats on the Inc 5000 listing") but not a buying signal.

• Conference speaking/sponsoring: Shows where they invest attention and marketing budget.

• New customer wins (theirs): Indicates growth and potentially new requirements.

Low-Signal / Noise (skip unless directly relevant)

• Generic press releases: Product updates, seasonal announcements, corporate boilerplate.

• Industry analyst mentions: Usually too abstract to be actionable.

• Social media corporate posts: Rarely reveals anything useful for sales intelligence.

Tech Stack Detection

Understanding a prospect's technology stack tells you about their sophistication, budget, and potential integration needs. Here are the best detection methods:

Job Postings (highest value)

Job descriptions are the most honest source of tech stack data. Companies list the actual tools they use because they need candidates who know them.

• Search: "[Company Name]" jobs OR careers [tool category]

• Look for: "Experience with [tool]", "Proficiency in [technology]", "We use [platform]"

• Engineering posts reveal development stack. Marketing posts reveal marketing tools. Sales posts reveal sales stack.

Website Analysis

• View the website source or use technology detection services (BuiltWith, Wappalyzer)

• Look for: analytics tags (Google Analytics, Segment, Mixpanel), chat widgets (Intercom, Drift), CMS indicators (WordPress, Webflow), CDN (Cloudflare, Fastly)

• Pricing page technology indicators: Stripe billing, Chargebee, usage-based metering tools

Engineering Blog / Technical Content

• Many companies publish their stack decisions on their engineering blog

• Search: "[Company Name]" engineering blog OR tech stack OR architecture

• Conference talks by their engineers often reveal stack choices

Integration Ecosystem

• If the company is a SaaS product, check their integrations page. The tools they integrate with are often the tools their customers (and they themselves) use.

Hiring Signal Analysis

Job postings reveal more about a company's priorities than almost any other public data source. Here is what different hiring patterns signal:

Volume Signals

• Hiring aggressively (10+ open roles): Growth mode. Likely has budget. May be experiencing scaling pain that ${company_name}'s solutions can address.

• Minimal hiring (0-2 roles): Either stable/profitable or constrained. Check other signals (funding, layoff news) for context.

• Hiring spree followed by pause: Possible pivot, budget tightening, or reorganization. Timing may be wrong.

Role Type Signals

• Hiring for ${company_name}'s product category: (e.g., hiring a role related to the solutions described in Organization Context) -- STRONG buying signal. They're investing in the function ${company_name}'s product serves.

• Hiring SDRs/AEs: Building out sales motion. May need sales tools, CRM, enablement.

• Hiring engineers: Building product. May need dev tools, infrastructure, CI/CD.

• Hiring a Head of [Function]: New leader = new strategy, tool review, vendor evaluation. Prime window.

Technology Signals

• Job descriptions that list competitor tools: "Experience with [Competitor] a plus" -- they use your competitor. Competitive displacement opportunity.

• Job descriptions that list ${company_name}'s product: "Experience with [${company_name}'s product]" -- they already use it (expansion) or are evaluating it.

Connection Point Discovery

The #1 goal of lead research is to find connection points that make outreach feel personal and relevant. Here are the best connection points, ranked by effectiveness:

Tier 1: Direct Relevance (reference in opening line)

• They posted about a problem you solve. "I saw your post about [X] -- we help teams like yours handle exactly that."

• They recently changed roles. "Congrats on the new role at [Company]. When I joined a new team, one of the first things I evaluated was [relevant category]."

• Their company just announced something relevant. "Saw the Series B announcement -- congrats. Companies at your stage often start looking at [your category]."

• Mutual connection. "I noticed you're connected with [Name] -- they've been using our product for [use case]."

Tier 2: Contextual Relevance (reference in body)

• They spoke at a conference on a topic related to ${company_name}'s product area.

• They published an article or blog post touching on your space.

• Their company is hiring for a role in ${company_name}'s product function area.

• They share a professional background (same company alumni, same university, same industry transition).

Tier 3: Light Personalization (use if nothing better)

• Company milestone (anniversary, growth achievement, award).

• Industry trend they're likely affected by.

• Geographic connection (same city, same region).

Rule of thumb: Always aim for at least one Tier 1 connection point. If you can't find one, you need to dig deeper or reconsider whether this lead is worth manual outreach (vs. an automated nurture sequence).

Data Freshness Standards

Data decays fast in B2B. Here are the freshness standards. See references/source-hierarchy.md for the complete data source reliability hierarchy, cross-reference rules, conflict resolution methodology, and provider-specific accuracy data.

Data Type Fresh Acceptable Stale Action on Stale

Job title / role <3 months 3-6 months

6 months Flag as "may have changed"

Company size <6 months 6-12 months

12 months Triangulate with other sources

Funding data <3 months 3-12 months

12 months Note date explicitly

Tech stack <6 months 6-12 months

12 months Cross-check with job postings

News / press <30 days 30-90 days

90 days Only include if highly significant

Contact details <3 months 3-6 months

6 months Verify before outreach

Always include the date of the most recent data source. A profile built on data from 2 weeks ago is dramatically more valuable than one built on data from 8 months ago.

Output Contract

Return a SkillResult with:

• data.lead_profile : Structured profile object with:

• name : Full name

• title : Current job title

• linkedin_url : LinkedIn profile URL (if found)

• role_seniority : "C-level" | "VP" | "Director" | "Manager" | "IC"

• tenure_current_role : How long in current role (if determinable)

• tenure_current_company : How long at current company (if determinable)

• background : Brief professional background summary (2-3 sentences covering career arc)

• previous_roles : Array of last 2-3 roles with company, title, and approximate dates

• recent_activity : Any recent posts, talks, publications, or public activity

• content_topics : Topics they post about or engage with on LinkedIn

• decision_authority : "likely_decision_maker" | "likely_influencer" | "likely_evaluator" | "unknown" (inferred from title + company size)

• data.company_overview : Company details with:

• company_name : Official company name

• website : Company website URL

• industry : Industry classification

• company_size : Employee count range

• headquarters : Location

• founded : Year founded (if found)

• description : One-paragraph company description

• business_model : How they make money (SaaS, services, marketplace, etc.)

• data.recent_news : Array of 3-5 recent news items (high-signal only) with:

• title : Article/news title

• source : Publication name

• date : Publication date

• summary : One-sentence summary

• relevance : Why this matters for outreach

• url : Link to the article

• data.enrichment_data : Additional intelligence with:

• funding : Latest funding round, amount, investors (if available)

• tech_stack : Known technologies and tools (with detection method noted)

• hiring_signals : Recent job postings and what they indicate for your sales motion

• growth_indicators : Revenue, headcount growth, market expansion signals

• data.connection_points : Array of 2-5 specific connection points for outreach, ranked by tier (1/2/3), each with:

• point : The connection point

• tier : 1, 2, or 3

• suggested_use : How to reference it in outreach

• source : Where you found it

• data.data_freshness : Date of the most recent data source used

• references : Array of source URLs used in research

Quality Checklist

Before returning research results, verify:

• Person identity is confirmed (not a name collision with someone at a different company)

• LinkedIn URL is verified as the correct person (not just a search results link)

• Company description is accurate and current (not a 3-year-old About page for a company that has since pivoted)

• At least one Tier 1 or Tier 2 connection point is identified (if none found, note this gap explicitly)

• News items are high-signal, not noise (no generic press releases or industry boilerplate)

• Tech stack data includes detection method (job posting vs. website analysis vs. integration page)

• Hiring signals are interpreted, not just listed ("Hiring 3 SDRs" is data; "Building out outbound sales motion -- may need sales enablement tools" is intelligence)

• All data points have a cited source URL

• Data freshness is noted for time-sensitive fields (title, company size, funding)

• Gaps are explicitly called out (what you looked for but could not find)

• Output is scannable -- a rep can get the key points in 30 seconds

Guidelines

• Always cite sources with URLs so the rep can verify information

• If the lead name is ambiguous, use company context to disambiguate. If still ambiguous, present the most likely match and note the uncertainty.

• Prioritize recent information (last 6 months) over older data

• If limited information is found, clearly state what could not be determined rather than guessing. "Not found" is always better than fabricated data.

• Use ${company_name} context to tailor research toward relevant competitive and partnership angles

• Flag any connection points between the lead's company and ${company_name} (shared investors, mutual connections, technology overlap)

• Focus on intelligence, not data. The rep doesn't need to know that the company was founded in 2015 in Austin. The rep needs to know that the company just raised $30M and is hiring 5 engineers, which means they're scaling and probably evaluating new tools.

Error Handling

No lead name or company name provided

Ask the user for clarification. Provide a template: "To research this lead, I need at minimum a person name and company name. If you have an email address or LinkedIn URL, those help me find the right person faster."

Ambiguous person (common name, no company context)

Do not guess. Return the top 2-3 possible matches with distinguishing details and ask: "I found multiple people named [Name]. Which of these is the right one?" If company is provided but person is ambiguous within the company, use title/department context to narrow down.

Company is very small, private, or new

Small companies have less public data. This is expected, not an error. Lean harder on:

• Founder/CEO LinkedIn profile (often the richest source for small companies)

• Product Hunt, AngelList, Crunchbase (startup databases)

• Social media presence (Twitter/X, LinkedIn company page)

• Website itself (especially the blog, which small companies often use as their primary content channel)

Note limited data availability honestly: "Limited public data available -- [Company] appears to be an early-stage startup with fewer than 20 employees. Profile confidence is medium."

Web search returns no results for the person

The person may have a minimal online presence. This itself is information:

• Return what you found about the company

• Note: "No significant online presence found for [Name]. This may indicate they are new to the role, prefer a low profile, or the name may be different on professional platforms."

• Suggest alternative approaches: "If you have their LinkedIn URL or email, those would help me confirm their identity."

Email domain is a personal email (gmail, yahoo, etc.)

Personal email provides no company context. Flag this: "Email is a personal address -- cannot determine company affiliation from the domain. To research this lead, I need a company name or LinkedIn profile." If a name is available, search for the name directly, but note the lower confidence.

Data sources conflict

When sources disagree (e.g., LinkedIn says 200 employees, Crunchbase says 150):

• Present both data points with their sources

• Use the most recently updated source as the primary

• Note the discrepancy: "Employee count varies by source: LinkedIn company page shows ~200, Crunchbase shows 150 (updated 3 months ago). Using LinkedIn figure as primary."