12 Claude Prompts for B2B Customer Success Managers
CSMs ship the same five reports every Monday and the same three save-call briefs every Friday. The prompts below cut that to a 13-second copy-paste each. All twelve are written for Claude with the LinkFetch MCP attached. Wire them to a daily, weekly, or trigger cadence and you've replaced a junior analyst.
LinkFetch, a compliance-first LinkedIn data API, is the source of truth for every external signal: profile changes, post activity, hiring deltas, public sentiment. None of these prompts scrape, none of them touch logged-in data, and all of them run on the same flat per-call pricing the rest of the LinkFetch library uses.
How to actually use these prompts
The biggest mistake CSMs make with AI tooling is treating each prompt as a standalone chat. The ones below are designed to run on a schedule, output structured data, and feed your existing tools (Vitally, Gainsight, ChurnZero, Salesforce). Define the output format upfront: tell Claude exactly what shape you want back, then pipe it to wherever you already triage your book of business (Vitally, 2026).
Three setup steps before any prompt runs:
- Wire LinkFetch's MCP to Claude Desktop. The standard config is in the daily outreach recipe.
- Upload your champion list, account map, and ICP definition once. Claude remembers them across the session and the prompts reference them implicitly.
- Set a cadence per prompt in your scheduler of choice. The
cadencefield in this post's frontmatter is the default we recommend.
The twelve prompts below split into four buckets: champion tracking (daily and weekly), expansion mining (weekly), risk detection (weekly), and trigger-based briefs (one-shot). Run the daily ones every morning, the weekly ones on the day labeled, and the triggers when the named event fires.
The four prompts that pay for the seat in week one
If you can only run four of the twelve, run these. They cover the 80% of CSM workflows that move the needle on retention and expansion this quarter.
1. Daily champion-change alert. The single highest-ROI prompt in the library. Champion Follow and similar tools track key customer contacts and notify CSMs in real time when they leave or change jobs (Aircover, 2026). The Claude version costs roughly 30 credits a week and catches changes the next morning, which is fast enough for any non-enterprise renewal cycle.
2. Renewal risk early warning. A weekly Friday pull that ranks your book by two leading indicators: champion silence and unfilled seats around them. Most renewal-risk dashboards fire when the problem is already a problem. This one fires three to six weeks earlier because it watches signals upstream of usage drop.
3. QBR prep pack. Most CSMs spend three hours pulling stakeholder context the night before a QBR. This prompt does it in one Claude turn, with the stakeholder's exact words quoted from their public posts. The slide where you reference their last LinkedIn post about their priority is the slide that gets you the renewal.
4. Multi-threading gap report. The accounts that churn are the accounts where you knew one person. This weekly Monday pull compares your CRM contacts against the real org chart on LinkedIn and flags every account where the math says "you're one resignation from a churn event."
The eight that compound month over month
The remaining eight prompts in the frontmatter cover the long tail: champion promotion congrats, competitive-loss scanning, reference sourcing for sales, cross-sell intro mining, save-call prep, and the monthly champion network map. None of them are urgent in any given week. All of them shift retention and net revenue retention by a percentage point or two over a quarter, which on a book of $5M ARR is the cost of two CSMs.
The monthly champion network map is the underrated one. It uses your 30 strongest customer relationships as a referral source, fitlered through your ICP, and outputs a list of 50 warm prospects ranked by who can intro. Most CSMs never run this and most CS leaders never ask for it. It's a referral pipeline hiding in your book of business.
Why LinkedIn is the right data spine for CS in 2026
Two independent 2026 studies confirm LinkedIn is now the second most cited source across ChatGPT, Perplexity, Gemini, and Claude when those models answer business questions (ALM Corp, 2026). That citation density means LinkedIn data is what your customers' AI agents are reasoning from too. Building your champion graph, your account map, and your competitive intelligence on the same data source they read keeps you aligned with the worldview your customers operate inside.
LinkFetch is the API layer that makes this affordable. Profile checks at flat per-call pricing, no per-result math. Company data including hiring activity. None of it requires a logged-in scraper or a Sales Navigator seat per CSM.
How this compares to other CSM prompt libraries
Vitally publishes a strong 23-prompt library focused on call prep, data analysis, and churn diagnosis (Vitally, 2026). Their prompts are tool-agnostic and assume the CSM pastes data in manually. The 12 above are tool-aware and assume LinkFetch is wired to Claude through MCP, which is the difference between a prompt you run once and a routine you forget is running.
For the AE side of the same workflows, see the AE prompt library, which covers pipeline, opportunity, and forecast prompts. For the RevOps buyer of these same routines, the sales-ops prompt library maps the same loops to forecast and territory work. The churn-watch recipe is the long-form version of prompt 1 here, with the full Claude Desktop config and a 60-day case study.
FAQ
Will running 12 scheduled prompts blow my LinkFetch credit budget?
No. The full library, run on the cadences shown in the frontmatter for a 200-account book, costs roughly 400 credits a week. That's under the cost of a single Sales Navigator seat at flat per-call pricing, and an order of magnitude under what a dedicated CS Ops analyst would cost to produce the same outputs.
Can I run these in Gainsight or Vitally instead of Claude Desktop?
The Claude Desktop MCP integration is the cleanest path because the LinkFetch tool surface is native. Gainsight and Vitally both support Anthropic API integrations, so you can move the same prompts there once you've validated them. We recommend running the first 30 days in Claude Desktop and porting the proven prompts afterward.
What if my champion list is 500 contacts, not 30?
The daily champion alert scales linearly: 500 contacts checked daily is roughly 500 profile calls a week, or 100 credits a day on flat per-call pricing. That's still under one CSM-hour a week in spend. The monthly network map is the only prompt where you should trim, since the network mining is per-contact expensive. Cap that one at your top 30.
How fast does the daily champion alert catch a job change?
Within 24 hours of the LinkedIn profile being updated by the champion. LinkFetch refreshes the profile on every call, so the moment a champion edits their headline, the next morning's run catches it. Champions usually update LinkedIn within 7 days of starting a new role, which gives the CSM a one to two-week head start on the standard renewal-risk dashboard.
Last updated 2026-06-22 by the LinkFetch team.