Prompt

Find accounts posting about the problem you solve

Tools: Claude ▪ Lusha

Outputs on this page come from a live run on September 10, 2026. Company names and post authors are masked. No contacts were previewed or revealed, and the people mentioned in posts are not listed.

Of the 24 company signals in the Lusha plugin for Claude, this is the newest and the one your team asked for most. It classifies what employees post into intent categories, returns the author as metadata, and refreshes daily. Used well, it tells you who to reply to. Used as a filter, it returns everyone.

The prompt

Open in Claude Open in ChatGPT

Find which of my target accounts have employees posting on LinkedIn about the problem I solve, and tell me which posts are worth acting on.

My topics (the words that would appear in a post from someone who needs what I sell): [e.g. sales hiring, SDR, sales development, lead management, sales operations, revenue operations, prospecting, contact data]

Before anything else, run account_usage and tell me my remaining credits. Do not run any signal call if fewer than 60 credits remain; stop and tell me.

Here is my list (domain, one per line):
[PASTE UP TO 25 DOMAINS]

Customer domains (or write "none"):
[PASTE]

Steps:
1. Run signals_company_filters with filterType "intentCategories" and pick the 2 or 3 categories closest to my topics. Tell me which you chose. These categories are broad, so they are a net, not a filter.
2. Run signals_companies_search on the list with signalTypes ["linkedinActivityIntent"], filters.include.intentCategories set to those categories, startDate 30 days ago, maxResultsPerSignal 2. Each post returned costs 1 credit, so keep the window short. Batch in groups of 25 if the list is longer.
3. For every post, pull the publication date, the signalCategory (Pain Point or Challenge, Technology or Stack Discussion, Thought Leadership, Company or Career Update, Event or Conference Reference, Community or Recognition), the activity type (post or repost), the matched topic names, the one line summary, and the engagement counts.
4. Keep only posts where at least one matched topic name contains one of my topic words. A category match alone is not enough; "CRM" as a category tags posts about customer satisfaction at insurance brokers. Drop reposts and drop Event and Recognition posts unless a topic word matches directly.
5. Rank what is left: Pain Point or Challenge first, then Technology or Stack Discussion, then Company or Career Update where the topic is hiring for a role I sell to, then Thought Leadership. Within a tier, rank by engagement.
6. For each kept post, describe the author by role only (for example "sales manager", "recruiter", "CEO"). Flag the author for a title check before any outreach: attribution can lag, and some authors no longer work at the company.
7. Tag customers as EXPANSION and everyone else as NEW LOGO. For the top 5 NEW LOGO posts, write one opening line that responds to what the post is about without quoting it and without mentioning that you saw it on LinkedIn.

Output: one table with columns Account | Post type | Matched topic | Date | Author role | Engagement | Segment | Opening line. Then three lines: how many posts were returned, how many survived the topic filter, and credits used. Do not preview or reveal any contacts in this run, and do not list the people mentioned in the posts.

What you get

The situation. A list of 25 mid market B2B software companies, 800 to 8,000 employees. Topics: sales hiring, SDR, sales development, sales operations, revenue operations, lead management. Categories chosen: Sales, Demand Generation, CRM. No customer domains supplied.

The run. account_usage first, at zero credits. Then one batch call on 25 domains for linkedinActivityIntent with the three categories, 90 day window, two posts per account. 24 of 25 fired, 48 posts, 48 credits. Every post was from the last four weeks. After the topic word filter and dropping reposts, events and recognition posts: 9 posts at 8 companies.

What came back.

Account	Post type	Matched topic	Date	Author role	Engagement	Segment	Opening line
K.L. (marketing platform, ~2,000 emp)	Company update	Sales growth, revenue growth management	Sep 4	Senior executive	87 likes	NEW LOGO	Standing up a regional headquarters at that scale means a lot of new reps and new territories at once. The first thing they need is a book that is already verified.
T.W. (communications platform, ~5,000 emp)	Company update, hiring	Sales hiring, sales enablement	Sep 5	Sales manager, EMEA	15 likes	NEW LOGO	New SDRs usually spend their first fortnight finding phone numbers instead of dialing them. Happy to show how EMEA teams skip that step.
C.K. (cybersecurity, ~7,000 emp)	Company update, hiring	Sales development, sales hiring	Sep 4	Recruiter	19 likes	NEW LOGO	Building an SDR team for a new division from graduates means the data has to do the qualifying the reps cannot yet. That is a solvable problem before day one.
J.F. (DevOps platform, ~1,500 emp)	Company update, new role	Sales management, sales coaching, sales hiring	Aug 3	New sales development leader, APAC	94 likes	NEW LOGO	Congratulations on the APAC role. Building an SD function in a new region starts with knowing which accounts are real; that is the part we can take off the list.
D.B. (file storage, ~2,500 emp)	Company update, hiring	Sales strategy	Sep 4, Aug 31	Two employees, same job post	0	NEW LOGO	Hiring a sales strategy and operations lead for APJ says the region is getting structure. Worth a conversation about what the data layer under that structure looks like.
S.I. (analytics, ~500 emp)	Thought leadership	Sales operations	Aug 28	Sales operations lead	82 likes	NEW LOGO
W.X. (website platform, ~5,000 emp)	Company update, hiring; event	Sales development, revenue intelligence	Sep 3, Aug 24	Sales leader; marketing	4, 10 likes	NEW LOGO
O.K. (identity, ~6,000 emp)	Company update, hiring	Sales hiring (finance role)	Sep 4	Finance manager	8 likes	NEW LOGO	Kept by the filter, but the role is a revenue accountant. Low fit.
39 posts at 20 companies	Dropped	Category match only: customer satisfaction, Salesforce developer, product manager, event recaps, awards, culture posts, reposts
1 account	no signal

48 posts returned, 9 survived the topic filter, 8 companies. Credits used: 48. Confirmed live via Lusha connector, September 10, 2026. Names masked.

What happens next. The author is the person to reach, because they are already talking about the problem in public and will answer a relevant reply. Hand each author to the verify a contact’s title play first: in this run at least two posts were attributed to a company the author appeared to have left. Then reply where they posted, or reach out through the verified channel; the opening lines above are written for either. The four hiring posts also confirm what Surge in Hiring by Department reports from job boards, which is a stacked signal for the price of one more call.

Why use Lusha in Claude

Every other signal in the set infers intent from something a company did: hired, spent, raised, changed leadership. This one records something a person said, and the person is named. That is why it converts when it is right and why it costs so much when it is wrong; you are paying per post, and people post about everything. Inside Claude, the plugin fetches the posts for the whole list in one call and the prompt does what a rep would do with an afternoon and a feed: reads the type of post, checks whether the topic is actually the topic, ignores the reposts and the awards, and keeps the handful where someone said something you can answer. Every signal is sourced under Lusha’s data methodology and compliance framework, and the author’s contact details are never touched in this play.

FAQ

What counts as LinkedIn activity intent?

A public LinkedIn post or repost by an employee of the company, classified into topic categories and a post type, with the author returned as metadata. The signal is company level; the author is context.

Why filter by my own words after filtering by category?

Because the categories are broad by design. “Sales” covers product managers and “CRM” covers customer satisfaction anywhere. Your words are the filter; the category is the net that makes the fetch affordable.

Which post types carry intent?

Pain Point or Challenge, when it appears, is the strongest and the rarest: one in 48 in the live run. Technology or Stack Discussion is next. Career updates and job posts are strong only when the role posted is one you sell to. Events, awards and culture posts almost never are.

Can I contact the author?

Yes, through the standard verified contact flow, after confirming their current title. The post tells you what to say; it does not give you their email, and the prompt does not reveal one. Do not reference the post itself; reply to the topic.

Does this spend credits?

One credit per post returned, before any filtering. In the live run, 25 domains over 90 days cost 48 credits for 9 usable posts. A 30 day window on the same list would have cost roughly a third. account_usage runs first and the prompt stops under 60. No contact preview or reveal runs in this play.

How often should I run it?

Weekly, with a 30 day window, on a list of accounts you already work. It is a conversation starter for accounts in play, not a list builder.

Ready to run this?

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

Start for free Read the Docs