Skip to Content
CoursesMCP courseAccount research

Module 06 · Research a company

Outcome: an account brief — who they are, what they run, where their traffic comes from, what they are shipping and what they are advertising — assembled in one message for under 15 credits.

Surface
MCP server
Level
Beginner
Uses
find_companiesenrich_organizationenrich_linkedin_pagecompany_product_launchfind_company_techstackfind_company_website_trafficsemrush_insightssimilarweb_insightsb2b_ads_searchmeta_ads_analysis
Credits
~13.5 to run every example once, at limit 10 on the search
Prerequisite
A company domain you know

Tools in this module

ToolCostRequired inputReturns
find_companies0.3 per resultNone — pass a filterMatching company profiles
enrich_linkedin_page0.5identifierDescription, specialties, employee count
b2b_ads_search0.5company_domain or company_nameB2B ads with creative and targeting
meta_ads_analysis0.5page_idActive Meta ads across all countries
company_product_launch1domainRecent launches with dates
find_company_techstack1domainTechnologies in use
find_company_website_traffic1domainTraffic, cheapest option
enrich_organization2domainIndustry, size, funding, technologies
semrush_insights2domainTraffic, organic keywords, backlinks
similarweb_insights2domains (array)Visits, engagement, sources, competitors

Three tools return traffic at three prices. find_company_website_traffic at 1 credit is a provider-agnostic waterfall and is the right default. Pay 2 for SemRush when you want keywords and backlinks, or SimilarWeb when you want traffic sources and competitors.


Example prompts

find_companies — build the account list

Filter families: identity, category, location, type, size, revenue, funding, social, age and growth. Full table.

Use find_companies with: - industries ["Computer Software"] - exclude_industries ["Staffing and Recruiting"] - countries ["United States"] - company_types ["Privately Held"] - employee_count_ranges ["51-200","201-500"] - min_total_funding_usd 5000000 - last_funding_round_types ["series_a","series_b"] - last_fundraise_after "2025-06-01" - min_headcount_growth_percent_6m 20 - min_year_founded 2018 - limit 25 Return: company_name, primary_domain, employee_count_range, countries. Stop after one page and report credits used.

Dates are YYYY-MM-DD. “Last six months” will not parse.

enrich_organization — the firmographic record

Use enrich_organization with domain "stripe.com". Return industry, employee count, HQ location, total funding, last round and the technologies it reports.

enrich_linkedin_page — the cheap version

Use enrich_linkedin_page with identifier "https://linkedin.com/company/stripe". Return description, specialties and employee count.

enrich_linkedin_page is 0.5 credits against enrich_organization at 2. If you only need the description, size and specialties, take the cheap one. identifier accepts a LinkedIn id, a full URL or a company name.

company_product_launch — what they are shipping

The details parameter narrows the search — use it, or you get everything.

Use company_product_launch with domain "stripe.com", details "AI and automation product launches in the last 12 months". Return launch name, date and a one-line summary.

find_company_techstack — what they run

Use find_company_techstack with domain "stripe.com". Group the result by category and flag anything in the CRM, sales engagement or data enrichment categories.

find_company_website_traffic — the 1-credit traffic read

Use find_company_website_traffic with domain "stripe.com". Return monthly visits and the trend direction.
Use semrush_insights with domain "stripe.com". Return estimated traffic, the top 20 organic keywords by position, and total referring domains.

One domain per call.

similarweb_insights — sources and competitors

Use similarweb_insights with domains ["stripe.com","adyen.com"]. Return total visits, bounce rate, traffic sources by channel, and the listed competitors for each.

Takes an array — compare an account against its rival in one call.

b2b_ads_search — what they are saying in paid

Use b2b_ads_search with company_domain "stripe.com". Return each ad's headline, body and targeting. Tell me which pain point they lead with.

Prefer company_domain. company_name has to resolve first and ambiguous names fail.

meta_ads_analysis — the Meta side

Use meta_ads_analysis with page_id "https://facebook.com/StripeHQ", ad_type "all ads", max_results 20. Return creative, copy and how long each ad has been running.

max_results accepts 10–20 here, not 100.


The 5.5-credit account brief

Identify

enrich_linkedin_page at 0.5 — description, specialties, size.

Qualify

find_company_techstack at 1 — do they run anything you integrate with or replace?

Size the demand

find_company_website_traffic at 1 — the cheap traffic read, not SemRush.

Find the hook

company_product_launch at 1 and b2b_ads_search at 0.5 — what they shipped, what they are promoting.

Brief me on stripe.com. In this order: 1. enrich_linkedin_page with identifier "stripe.com" 2. find_company_techstack with domain "stripe.com" 3. find_company_website_traffic with domain "stripe.com" 4. company_product_launch with domain "stripe.com", details "last 12 months" 5. b2b_ads_search with company_domain "stripe.com" Return one brief under 200 words: what they do, what they run, how big their traffic is, what they just shipped, what they advertise. Then one sentence on why they would care about us. Report credits used.

Swap in enrich_organization and semrush_insights when the account is worth 4 more credits.


Workflow prompts

An account brief is only useful where the account already lives. These write it back.

Brief every new account the moment it lands in the CRM

  • pull accounts created this week with no research on them
  • build the brief
  • write it to the record
1. Pull HubSpot companies created in the last 7 days where the "Tech stack" property is empty. Give me the count first. 2. For each domain, in this order: enrich_linkedin_page → find_company_techstack → find_company_website_traffic → company_product_launch with details "last 12 months". 3. Show me one row per account: company, size, stack highlights, traffic, latest launch. Roughly 3.5 credits each — confirm the total with me. 4. On my go-ahead, write the stack to the "Tech stack" property and the brief to a note on the company record. Do not overwrite existing values.

Score an inbound list against your ICP before anyone works it

  • take the raw list
  • pull the qualifying signals only
  • grade and rank, spending nothing on the rejects
For each domain in this list: 1. enrich_linkedin_page — 0.5 credits — for size and description. 2. STOP on any company under 20 employees or in staffing/recruiting. Mark them DQ and spend nothing more on them. 3. For survivors only: find_company_techstack and find_company_website_traffic. 4. Grade A/B/C on: headcount 51-1000, runs a CRM we integrate with, traffic over 20k monthly. 5. Return the table sorted A first, and tell me how many you DQ'd at step 2 and what that saved.

Competitive brief for a live deal

  • compare the account against its closest rival in one call
  • pull what both are advertising
  • hand back the talk track
For a deal against <competitor>: 1. similarweb_insights with domains ["<prospect domain>", "<competitor domain>"] — one call, both companies. 2. b2b_ads_search with company_domain "<competitor domain>". 3. company_product_launch with domain "<competitor domain>", details "last 6 months". Give me: where the prospect's traffic comes from versus the competitor's, what the competitor leads with in paid, and what they have shipped recently. Then three talking points for the call. Under 200 words.

Best prompting practices

  • Search before you enrich. 100 companies from find_companies is 30 credits. Guessing 100 domains and enriching each at 2 is 200.
  • Use the range enums. employee_count_ranges takes fixed buckets — "51-200", "1001-5000". “Mid-market” is not a value.
  • Exclude early. exclude_industries costs nothing and keeps agencies and recruiters out of a list you would clean by hand.
  • Pick one traffic tool. Running all three on the same domain is 5 credits for one number.
  • Pass details to company_product_launch. Without it you get every announcement the company ever made.
  • Use domains over names wherever a tool accepts both.

Where this breaks

find_companies bills 0.3 credits per company returned, so paging is what drives spend. Three pages of 100 is 300 companies and 90 credits. State how many you want up front.

meta_ads_analysis needs a Facebook page id or page URL, not a domain. If you only have a domain, get the page from enrich_organization first.


Further automation

MCP Account Brief  is this module as a skill file — /mcp-account-brief runs the cheap path by default and names the reason before any upgrade.

Account Researcher and ICP Qualifier run it across a list of domains and return one scored row per account.

Browse all GTM agents →


Next module

06 — Growth and hiring signals, the ten tools that tell you an account is spending money right now.

Reference: Find Companies · Enrich Organization · Enrich LinkedIn Page · Product Launch · Tech Stack · Website Traffic · SemRush · SimilarWeb · B2B Ads · Meta Ads · TAM Sourcing