Prompt

Clean up prospecting lists using decline signals

Tools: ClaudeLusha

Prospecting lists grow. They almost never shrink on their own. An account added six months ago because it showed growth or funding can still be sitting on a rep’s list today, long after that signal reversed — because nobody checks a list for the absence of momentum, only its presence. Connect Lusha to Claude and this prompt checks every account on a list for any decline signal, and returns a shorter, cleaner list of who’s actually still worth working.

The prompt

 

Here is my current prospecting list: [paste company names or domains].

Using the Lusha connector, check each account for any decline signal
in the last 90 days: headcount decrease, IT spend decrease, or website
traffic decrease. Return a table with company name, which decline
signal fired, and the date.

Then sort into two groups:
(1) no decline signal found - keep working
(2) one or more decline signals found - deprioritize or remove

Finish with the cleaned-up list of accounts still worth a rep's time,
and a separate short list of removed accounts with the reason each
was cut.

What you’ll get back

 

The situation: a rep pastes a 40-account prospecting list built up over the last two quarters and asks for a decline check before the next sprint.

The output:

Example output

Checked 40 accounts: 9 show a decline signal in the last 90 days

Kept (31): no decline signal found — still worth working, list unchanged.

Removed (9): [Company Q] — IT spend decrease, 34% drop, 45 days ago. [Company R] — headcount decrease, 3-month window, 12% drop. Seven more with the same structure, each with the specific signal and date.

Result: a 31-account list, down from 40 — every account still on it has shown no sign of pulling back in the last quarter.

Example outputs in this play are illustrative — they reflect the structure, fields, and format of real Lusha connector output, but were not pulled from a live session. Run the prompt with your own data and connectors to see live results.

Why it works

 

A prospecting list is usually built once, off a positive signal, and never re-checked. This prompt runs the check that normally never happens: not “who’s showing growth,” but “who’s stopped.” Three decline signal types run in a single pass, so an account doesn’t need to show every kind of pullback to get flagged, one is enough. The output is deliberately two lists, not one filtered list, so nothing gets silently dropped without a reason attached. Contact and company data is verified across multiple sources and handled under GDPR, CCPA, SOC 2 Type II, and ISO 27701 compliance.

FAQ

  • Does one decline signal mean I should drop the account entirely?

    Not necessarily — that’s a judgment call, not a rule. Some teams remove on any single signal, others wait for two. The prompt sorts accounts into “flagged” and “clean,” the decision to deprioritize versus fully remove is yours to apply.

     

     

     

     

  • How is this different from checking one specific decline signal, like headcount decrease?

    This runs all three decline types at once, so an account pulling back on IT spend but not headcount still gets caught. A single-signal check misses that.

  • How often should I run this on a prospecting list?

    Monthly is a reasonable cadence for a working list. Run it more often on a smaller, higher-priority list where stale accounts cost more in wasted rep time.

  • What if an account shows a decline signal but I still want to work it?

    That’s fine — the prompt flags, it doesn’t decide for you. An account with strategic value or an existing relationship might stay on the list despite a decline signal; the flag just makes that a deliberate choice instead of an oversight.

  • Is checking for decline signals like this compliant?

    The prompt reads business-level signals — headcount, spend, and traffic trend — the same category of public business information covered elsewhere in Lusha’s signal set. Lusha’s data is professionally sourced and opt-outs are honored automatically.

Ready to run this?

One data connection. Works in Claude, ChatGPT, your CRM, or any agent you build.