12 Claude Prompts for Growth Marketers
These are the prompts that move the most time on a growth marketing desk in 2026. ABM list construction, intent capture, paid-ads audience builds, content partnership scoping, and the weekly trigger scan that feeds the SDR queue. Run them through Claude Desktop with the LinkFetch MCP server connected, drop them into a Claude Project, or paste them one-off into chat when an account team comes asking for a brief. LinkFetch, a compliance-first LinkedIn data API, is the data layer; Claude does the orchestration and the prose.
How to read this library
Three columns matter when you sequence these into a stack.
The cadence is what decides whether a prompt becomes a Claude scheduled task, a manual weekly run, or a one-off rep request. Real-time prompts (none in this library, on purpose) fire on webhooks. Weekly ones go on a Monday or Tuesday morning batch and land in someone's inbox before the daily standup. On-demand prompts are the ones an AE or campaign manager pings you for, mid-week.
The credit estimate is the load-bearing number when you defend the budget. A 4-person growth team running the full library lands around 2,500 to 3,500 credits per month. That is the number to pre-commit on a LinkFetch monthly plan; from there the math becomes predictable. There is no cache discount and no overage multiplier on LinkFetch, so the budget number you set is the budget number you spend, give or take 10% on watchlist size.
The tools column is what Claude needs to reach via MCP. The
@linkfetch/mcp server exposes linkfetch.profiles,
linkfetch.companies, and linkfetch.companies.timeseries by
default. If you have not wired the MCP server into Claude Desktop
yet, the daily outreach recipe
covers the config in three minutes.
The four growth workflows these prompts collapse into
Twelve prompts, four workflows. Knowing which prompts belong together is more useful than reading any single one in isolation.
Account selection and the ABM list. Prompts 1 and 12 are the build-and-refresh pair for the Tier-1 list. Prompt 1 runs once a quarter when the ICP definition gets a meaningful update. Prompt 12 runs monthly and is the one that keeps the list from rotting. Most ABM programs die from list drift; running prompt 12 on a fixed Monday cadence is the single highest-leverage thing a growth team can wire into their stack.
Trigger events and outbound feed. Prompts 3, 4, and 7 are the weekly intent capture layer. They feed the SDR queue with named accounts and named people who hit a signal in the last 7 to 30 days. The reply rate on signal-based outbound runs 4 to 8x cold-list sends. Prompt 7 in particular is the highest-leverage of the three; recently-hired VPs of Marketing buy new tools at 5 to 10x the rate of seated execs in their first 90 days.
Paid audience and partnership. Prompts 5 and 10 are the two paid-marketing prompts. Prompt 5 builds a LinkedIn Matched Audience of 4,000 to 6,000 ICP-fit profiles for an ABM ads campaign. Prompt 10 surfaces creator-style profiles in the category who could co-amplify a piece of content. Both run monthly. The list from prompt 5 deduplicates against the prior month's upload so the audience refreshes without bloating.
Pipeline and post-event triage. Prompts 6, 8, 9, and 11 are the "after the event happens" prompts. A webinar fires, a competitor loses a VP, a quarterly review needs case-study candidates. These are the prompts a growth lead reaches for on demand. They typically run at a higher credit cost per fire because they touch larger lists, but they fire less frequently.
Walkthrough: prompt 7, the freshly-hired VP scanner
Prompt 7 is the one most growth teams adopt first. It is also the one that benefits the most from a careful read before deploying.
The cadence is weekly, not daily. Daily catches one or two profiles that fired in the last 24 hours, which is faster than any AE can actually act on. Weekly batches the last 7 days of new VPs into a single Tuesday morning list, which is small enough to read end-to-end and large enough to be worth a meeting. If your team is sending more than 8 outbound sequences per week against the output, move the cadence up. If they are sending fewer than 3, the prompt is not the bottleneck, the activation is.
The "between 7 and 60 days ago" window is deliberate. Anything under 7 days is too noisy (new hires that did not stick, contractor roles miscategorised, profiles that have not finished the LinkedIn role edit). Anything over 60 days is a soft cap on the new-exec buying-window. The internal buying behaviour data here is consistent: new execs are 5 to 10x more likely to evaluate new vendors in their first 90 days. The 7 to 60 day window captures most of that signal while filtering out the short-burst noise.
The company-size filter (100 to 1,500 employees) is where you tune to your ICP. Below 100 employees, the "VP Marketing" title is often a one-person marketing team, which is a different buying motion. Above 1,500 employees, the new VP usually inherits an existing vendor stack and a procurement process, which is also a different motion. Adjust the band to your ACV: lower ACV plays well below 100 employees with the founder or growth lead, higher ACV plays well above 1,500 with the new VP and their direct reports.
The output spec deliberately includes the exact start date when listed. The start date is the field the AE actually uses to time the outreach: prospects respond best when you reach out in week 3 or 4 of their new role (they have settled in, but the budget season for the year is still open). The "5 to 60 days ago" filter is for the prompt's pull, the "week 3 or 4" rule is for the AE's send timing.
Walkthrough: prompt 7's credit math and why it stays cheap
The credit math is the boring number that decides whether a prompt gets shipped. Prompt 7 looks expensive at ~100 credits per week, but the math works out cheaper than every alternative on the market.
A 100-credit weekly run produces 25 ranked profiles. At LinkFetch's flat per-request pricing, that is roughly the cost of a quarter of one Sales Navigator seat per year, for a workflow that Sales Navigator cannot run (it does not surface a chronological "new VPs this week" view, the closest equivalent is the "Recent Job Changes" filter which is noisier and slower).
Compared to manually scrolling LinkedIn's "started a new position" feed, prompt 7 takes a workflow that costs an AE 45 minutes per Monday and reduces it to a Tuesday morning email with the work done. The AE's first hour back at their desk is the most expensive hour of their week. Spending 100 credits to reclaim it is one of the cleaner ROI conversations a growth ops lead will have all year.
What to skip if you are starting from zero
If you are deploying the LinkFetch + Claude stack for the first time, do not run all 12 prompts at once. The library is meant to seed a quarter of work, not a week of it.
Start with prompts 3 (weekly trigger scan), 7 (freshly hired VPs), and 12 (monthly Tier-1 refresh). Those three cover the highest- leverage cadence: a fresh outbound feed every week, a fresh exec target list every week, and a fresh top-of-funnel list every month. Three prompts, roughly 600 credits per month, no other dependencies. Once those are running cleanly, layer in 8 (competitor churn) and 5 (paid audience) for the next two months.
The temptation is to ship the full library week one and demonstrate "AI doing all the marketing work". The cost is that nobody on the team owns any single prompt's output, so all 12 outputs land in a shared Slack channel and get ignored. Three prompts shipped with a named owner per prompt outperforms 12 prompts shipped to nobody, by about an order of magnitude in actual pipeline created.
FAQ
Do these prompts need the LinkFetch MCP server, or do they work with another LinkedIn data provider?
The prompts are written to call the linkfetch.profiles,
linkfetch.companies, and linkfetch.companies.timeseries tools
exposed by the LinkFetch MCP server. Substituting another LinkedIn
data provider means rewriting the tool names and the response
shape every prompt expects. The structure is portable, the syntax
is not.
What is the credit cost of running the full library for a quarter?
A 4-person growth team running all 12 prompts at the stated cadence spends roughly 8,000 to 11,000 credits per quarter. That includes the monthly Tier-1 refresh, weekly trigger scans, and quarterly case-study run. Add 20% for ad-hoc on-demand prompts (pre-meeting briefs, one-off list scoring). The total is well inside the mid-tier LinkFetch monthly plan.
Which prompts produce GDPR-sensitive output?
All of them touch LinkedIn profile data, which sits inside the public-profile lawful-basis envelope under both GDPR and CCPA when collected through a user's own LinkedIn session. The output of prompt 5 (paid audience CSV for LinkedIn Ads upload) is the one to review with your DPO first, because uploading a list of named people to a third-party ad platform changes the processing purpose. The others stay inside the standard sales-and-marketing legal basis.
How do I customise the ICP rubric without breaking the prompts?
Edit the criteria in prompt 1 and prompt 12 in place. Six rubric points is the working ceiling beyond which Claude starts inventing signals to fit the rubric. If you have a seventh signal you really care about, swap it for one of the six rather than appending it. Version the prompt in your repo alongside the rest of your marketing ops config so a rubric change is a code review, not a quiet edit.
Last updated: 2026-05-18 · Author: LinkFetch team