gtm-prospect

name

gtm-prospect

description

Prospecting hub. Opens by asking your GOAL, then routes to 7 sub-skills - find new companies, lookalikes of best customers, rank my accounts, event prospecting, expansion radar, TAM builder, champion tracker (with live watcher). Iterative Crustdata search, FIT x TIMING x WARMTH scoring, opt-in enrichment, handoff to sheet/CRM/sequencer. Use for "build me a list", "find companies like <customer>", "who should I prospect", "rank my accounts", "who's hot in my book", "list for <conference>", "upsell targets", "how big is this market", "track champions who leave customers".

gtm-prospect

The prospecting hub. Always open with: "What's your goal?" and show the 7 sub-skills. Route, then run that sub-skill's recipe exactly. Every recipe below was live-tested - follow it, including the gotchas.

STEP 0 (always): read config/gtm-config.md (ICP, what we sell, buyer titles, customer list, stack). Confirm the ask in one line before spending credits.

---

THE GOAL MENU (show this first)

1. 🎯 Find new companies - "I have an ICP, build me a fresh list"
2. 🏆 Lookalikes - "find companies like my best customers"
3. 📒 Rank my accounts - "here are my accounts (book / territory / CSV) - who do I work first?"
4. 🎪 Event prospecting - "who's at that I should meet?"
5. 📈 Expansion radar - "where can I grow inside my existing customers?"
6. 🌍 TAM builder - "how big is my market, with real numbers?"
7. 🔔 Champion tracker - "alert me when champions leave my customers" (live watcher)

---

SUB-SKILL RECIPES

1. Find new companies (net-new)

Intake: ICP (industry/size/geo/stage) + signals (palette below) + target titles + rows wanted.

1. Validate every categorical via autocomplete (linkedin_industries values like "Software Development"; country ISO-3 USA; stages snake_case series_a). Wrong value = silent zero.
2. company_search_db broad -> inspect top 10 -> refine 2-4 rounds (stage lock via last_funding_round_type, amount band, year_founded, growth floor employee_metrics.growth_12m_percent > N - filterable). Report what each round caught.
3. People pull: people_search_db, current_employers.company_id in [...] + titles. Junk filter (<100 conns, placeholder headlines, advisors/investors).
4. Score FIT x TIMING x WARMTH -> 🔥/🟡/⚪ with the signal + date shown. Free data only; enrichment opt-in.

2. Lookalikes

company_identify the seed customers (free, batch) -> company_search_db with company_id in seeds to read their shared traits (industry values, size band, stage, growth) -> those traits become the filters -> run recipe 1 from step 2. Exclude the seeds + existing customers.

3. Rank my accounts (book, territory, or pasted CSV - one sub-skill)

Input: account names/domains/CSV from anywhere.

1. Resolve ALL in one company_identify call (free, comma-separated, 25/batch). Flag unresolved rows honestly.
2. ONE company_search_db with company_id in [...], fields: funding date/stage/total, headcount. Gotcha (tested): employee_metrics.growth_*_percent often doesn't return via fields projection - drop fields (compact mode includes growth) or accept funding-recency as the main timing signal.
3. Optional depth per hot account: job_search with aggregations + limit:0 for open-role counts (cheap hiring signal).
4. Score: TIMING-weighted (funding <3mo = 🔥; 3-9mo = 🟡; >12mo = ⚪, growth/hiring bumps a tier). Output ranked table + "why" per row + this week's top 5.

4. Event prospecting

1. Source the roster for real - never invent attendees: web_search " sponsors exhibitors" -> the official /sponsors page usually lists tiers right in the search snippet (verified on a live conference) -> web_fetch the page for the full list. Rep-provided lists welcome.
2. company_identify the roster (free, batch) -> filter to ICP -> recipe 1 step 3 for the people (prefer folks posting about the event: social_posts_by_keyword on the event name).
3. Deliverable: meet-list ranked by ICP fit, with booth/tier + suggested opener referencing the event.

5. Expansion radar (make it WOW - this is revenue hiding in plain sight)

Input: customer list (CRM or rep). Per customer, sweep FOUR expansion surfaces:

• New money: raised recently / new round (company_search_db by id - funding fields)
• New people: exec hires in the function we sell to; team growth (people_search_db current_employers + start_date recent; job_search openings)
• New ground: teams/geos/departments we don't touch yet (people_search_db by function/region at the account vs where our contacts sit)
• Warm paths: our champions there + who they can intro (the referral ask, scripted) Output per customer: opportunity - surface - evidence (dated) - estimated size (seats/teams) - the warm path in - suggested play. Rank the whole book by expansion-readiness.

6. TAM builder (researched methodology - do it properly)

Bottom-up with real company counts beats top-down guessing (count actual companies x ACV). Build THREE layers, each = one company_search_db count query (limit:1, read total_count, ~3 credits each):

• TAM - broadest qualifying definition (anyone who could ever buy). Tested example: US software cos 50-1000 = 8,321.
• SAM - what our product/GTM serves today (add funding/stage/geo constraints). Tested: + raised $5M+ = 3,215.
• SOM - realistically winnable (SAM x credible win-rate %, or capacity: reps x deals/yr). Then: $ = counts x ACV (ACV from config or the rep). Cross-check top-down: web_search for analyst market-size figures and show both numbers side by side. Always state assumptions + the exact filters used (so it's defensible). Offer breakdowns (by size band / geo / stage) as extra count queries.

7. Champion tracker (list now + watcher forever)

1. The list today: people_search_db with past_employers.company_id in [customer ids] + recently_changed_jobs = true. Gotchas (tested): exclude advisors/investors/board (title filter (!) on "advisor|investor|partner" or drop them reading the list); use current_employers.name for the new company (the company_name key returns empty); junk-filter <100 conns.
2. The watcher (keeps it running): crustdata_watcher_create, event_type_slug: person-discovery-via-filters, event_filters: [{filter_type: LAST_COMPANY, type: in, value: [customer domains]}, {filter_type: RECENTLY_CHANGED_JOBS}] - RECENTLY_CHANGED_JOBS is boolean: NO type/value keys or it 400s (tested). LAST_COMPANY = the job they just left (do NOT use PAST_COMPANY - matches anyone who ever worked there). Creating is FREE; set frequency (7 = weekly) + expiration_date; no webhook needed (Crustdata-managed receiver; read hits via watcher_run_summary). Confirm with the rep before creating; surface the watcher id + the returned curl.
3. Output rows: Person - was [role] at [customer] - now [title] at [new company] - landed [date] - hand to gtm-outreach for the "congrats, you know us" touch.

---

SIGNAL PALETTE (offer during intake, composable)

Growth (headcount %, team growth, web traffic, revenue band) - Funding (recency, stage, size, investors) - Hiring (the role whose pain we solve, surge, tech in job posts) - Content & intent (people/company posting a keyword, competitor mentions) - People & movement (champion moved, new exec, competitor leavers, customer alumni, same school/YC) - Company events (new office, news, launch) - Tech & presence (competitor/category, G2/ProductHunt, Glassdoor, SEO) - Warmth (mutuals, shared investor, accelerator) - Disqualifiers OUT (layoffs, customers, competitors) - Custom AND-rules.

UNIVERSAL RULES

• Ask the goal first; confirm scope in one line; then run the recipe, narrating steps.
• Iterate, never dump round-1 results. Show what each refine round dropped and why.
• Free data first. Contact enrichment is opt-in + cost-confirmed ("emails for 20 = ~40-100 credits, go?"). Company enrich by company_id + exact_match:true only.
• Junk filter always: <100 connections, placeholder headlines, advisors/investors/board, geo/role mismatches.
• Watchers and CRM/sequencer writes need an explicit yes.
• Handoff (always ask): Google Sheet / CSV - push to CRM (confirm) - push to sequencer (enrich first, confirm) - hand to gtm-outreach - or table only.