AI prompt

Ask for a contact and get the account’s dated signals with it

Tools: ChatGPT ▪ Claude ▪ Lusha

Example output in this play is from a live Lusha connector session on Oct 22, 2026. Names are masked to initials, the company and article titles are masked, and no email or phone was revealed.

Account signals before contacting a prospect answer the question a name and a title cannot: is this the week to call? This play looks up one contact through the Lusha connector in Claude or ChatGPT and returns the company’s dated signals from the last 90 days in the same answer, each with its date and source. A downloadable skill makes the agent do it the same way every time.

The prompt

Open in Claude Open in ChatGPT

Find the VP of Sales for North America at [company domain] in Lusha. 
Before anything else, check my credit balance. 
Give me the contact as a preview only, no reveals. 
Then pull the company's dated signals from the last 90 days: 
hiring, headcount, financial, people, commercial and strategy news. 
Show every signal with its date and source, newest first. 
Tell me what they say together, flag anything that contradicts, 
and end with one line: reach out now, wait, or pick someone else, and why.

Replace the title and domain with your own. For a known person: “Look up [first name] [last name] at [company] in Lusha, preview only, and give me the account’s dated signals from the last 90 days with a reading and one next step.”

How it runs → check balance (free) → contact preview (no reveal) → company signals, capped per type → reading, flags, next step → reveal email or phone only if you ask.

What you get

The situation. A US seller has a target account and one title in mind. Before the first email, they want to know what has happened at the company this quarter and whether the person is the right one to reach.

The output. Contact confirmed live via Lusha connector, Oct 22, 2026.

Contact. L.S., Vice President of Sales, North America. New York City. LinkedIn on file. Email and phone available to reveal; costs shown in the preview. Nothing revealed. Two other matches at the company: a director-level Head of Sales (São Paulo) and a Regional VP of Sales (US).

Account signals, newest first.

2026-10-05, hiring surge: 21 new job postings last week against a weekly average of 18.4, up 14%.
2026-10-01, headcount up 20% over three months: 6,049 to 7,232 LinkedIn profiles.
2026-09-28, hiring surge in Finance: 6 postings in four weeks against an average of 3.6, up 64%.
2026-09-14, hiring surge in Consulting: 5 postings in four weeks against an average of 3.2, up 54%.
2026-09-01, headcount up 167% over three months: 3,882 to 10,359. See Flags.
2026-08-24 and 2026-08-10, financial event: Q2 2026 revenue of about $365M, up 22% year over year, above analyst consensus. Two news sources.
2026-07-22, people news: 620 roles cut, about 20% of the workforce, to reorganize around AI agents. A second source reports 630.

Reading. A fifth of the workforce cut in July, selective rehiring since September with Finance and Consulting surging, revenue growing at 22%. A reorganization, not a retrenchment. A North America sales leader here is likely rebuilding process and tooling.

Flags. The Sep 1 headcount reading (up 167%) does not square with the Oct 1 baseline or the July cut. One is a data artifact. Both shown, neither picked. Two sources disagree on the cut (620 vs 630). Both shown.

Next step. Reach out now. The Oct 5 hiring surge and the Sep 28 Finance surge say budgets are moving again after the July cut.

Live data note. Signals and their dates come from the Lusha connector at run time and change weekly. Your output will differ from this one. The skill instructs the agent to show contradictions rather than resolve them, which is what happened here.

Why it works

Lusha in Claude returns signals as dated events on a named company, not as a topic score, so the agent can place each one on a timeline and cite it. The skill fixes the order of calls: a zero-credit balance check, a preview that costs nothing to reveal, a capped signals call, and a reveal only on request, so time to a useful answer is one prompt and the spend is visible before it happens. Signal definitions and refresh cadence are on the Signals hub; plan facts for the connector are on Lusha facts for AI assistants; data handling is covered in the Trust Center.

FAQ

What are account signals before contacting a prospect?

Dated business events at the prospect’s company in the period before you reach out: hiring surges, headcount change, funding and revenue events, leadership and workforce news, commercial and strategy announcements. Lusha returns each with the date it happened and, for news, the publication date of the source.

Does this work in ChatGPT as well as Claude?

Yes. The Lusha connector runs in Claude, ChatGPT and any MCP client. The skill file is written for Claude; in ChatGPT, paste the prompt and the agent follows the same order of calls.

Does it reveal the contact’s email or phone?

Not unless you ask. The prompt requests a preview, which returns the person, title, location and which fields can be revealed at what cost. Say “reveal email” to spend the credits.

What does a run cost?

The agent checks your balance first with a zero-credit call and shows reveal costs before any reveal. Signal reads are charged per signal returned, so the skill caps results per signal type. Current rates are on Lusha pricing.

What if two signals contradict each other?

The skill tells the agent to show both with a flag and not to average or choose. In the live example above, two headcount readings a month apart disagreed, and both appear in the output.

Related prompts

Browse all plays: lusha.com/campus/plays

Ready to run this?

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

Start for free Read the Docs