Fill in what you know once — Section 1 builds the detailed company pull (full research tracker), Section 2 builds an optional wide-net bulk pull, Section 3 maps your target companies to your LinkedIn connections (and, optionally, your message history with them), Section 4 finds named department contacts via web search. v5
The main tool — full per-company research (up to 36 columns) and a fully formatted Excel tracker, built from the Shared Inputs above plus everything you set below.
A separate, cheap, wide-net prompt. Instead of full 31-column research per company, this asks Claude to build a fast flat list — Company, Website, Inferred NAICS Code, one-line description, Source — using the same competitor-cascade / analyst-list / conference-exhibitor / job-posting discovery methods as Section 1, just without the deep per-company research. Meant to be pruned by hand afterward: delete the rows that aren't relevant, then paste the survivors into "Company(s) you know" above and set Section 1's Discovery scope to "Just enrich these" so the full research pass only spends effort on companies you've already confirmed you want.
Since there's no per-company deep research here, this can go wider than Section 1's ~50-75 ceiling. Still bounded by how much a given seed/industry can actually support — Claude will say so if it can't find enough distinct real companies rather than padding the list.
This is a separate, focused prompt for cross-referencing your LinkedIn data export against your Top Targeted Companies to surface warm introduction paths. It always maps your target companies to your connections (Connections.csv). If you also attach your messages (messages.csv), it adds your conversation history with each of those connections, so you can see who you've already talked to, who's waiting on you, and who you've never contacted. Works best with 25 or fewer companies: the more companies you list, the more Claude has to fuzzy-match against for every single contact, which slows down the processing and can hurt match quality.
Run it separately from the Target Company Prompt Builder above: run that tool first, then paste your Top 25 companies below to build this prompt. Download the .txt, then paste that prompt into Claude and attach your LinkedIn export file(s) to the same message. There's no need to upload or paste the CSVs here — this tool never touches the actual files.
"Connections + Messages" (default) attaches both Connections.csv and messages.csv and gets you all four tabs below. "Connections only" attaches just Connections.csv and gets you the Summary and Contact Matches tabs. Use it if you don't have your messages file, or would rather not share your message history. Either way you still get the full company-to-connection map.
NOTE:
Steps to export your LinkedIn data (connections and messages):
What you get back (an Excel workbook, .xlsx):
If a target company has no matching contact in your CSV, it's still included as "No match found" rather than omitted, so you can see coverage across your full list. Note that LinkedIn's export only shows each connection's current employer, so "Former" is flagged only where the data says so (for example, "Ex EY" or a "Retired" title), and anything sent after your export date won't appear.
A separate, focused prompt that finds named individuals at a target company who work in or near a specific role/department — sourced from public web research (LinkedIn profiles, company team pages, press, case studies), not your own contacts. This is the fulfillment step for the main tracker's Key Contacts / Priority Titles column, which often comes back as a placeholder like "Practice-level title only — no named individual confirmed" — this prompt exists to turn that into real names. Works best with 15 or fewer companies, since each company/role pair needs its own live web search — a heavier per-company lift than Section 3's fuzzy-match.
Run it separately: paste your target companies below, along with the role/department you're targeting — the shortest variant from your tracker's Suggested Job Title Keywords column tends to work better here than the full title (e.g. "Cyber Risk OT" rather than "Managing Director, Cyber Risk OT" — dropping the seniority level widens the search instead of narrowing it to one level). Download the .txt, then paste into a new chat with web search enabled.
Pulled from your tracker's Suggested Job Title Keywords if you have one open — use the shortest variant. This is what the search matches against, tiered by closeness (see below).
"Yes" (default) also returns Tier 2 (adjacent practice, broader scope) and Tier 3 (umbrella/parent practice) matches when an exact match is thin — every result is still labeled with its tier, never blended in silently. "No" returns Tier 1 exact matches only, and says so plainly when a company has none, rather than substituting a looser match.
Match tiers used in this prompt:
For each match found, the report has Company, Matched Person's Name, Title, Department/Team (as stated in the source), Match Tier, Source, and a Confidence rating (High/Medium/Low). Companies with no usable public results are listed as "No public matches found" rather than dropped. The result is formatted as an Excel workbook (.xlsx), with rows ready to paste back into the tracker's Key Contacts / Priority Titles column.