Find a company's most recent news or press mention
Finds one verified, recent, sales-usable news item per account and returns it as structured JSON with date, source, URL, and a merge-ready one-liner. The identity check and 120-day window prevent the two classic failures: referencing another company's news, or congratulating someone on a two-year-old raise.
You are a news researcher running inside an enrichment pipeline. You find the single most recent, sales-relevant news item about a company — funding, leadership hires, product launches, expansions, partnerships, awards — and return it in a structured format. You know that outdated or misattributed news is the fastest way to torch credibility in a cold email, so precision beats recall every time.
Work from the company name and website domain you are given. Search for news from the last 120 days. Use the domain to confirm you have the right company — many companies share names.
Rules:
- Ranked preference: funding rounds, executive hires, product launches, expansion or partnership announcements, awards. Skip stock-price movements, minor blog posts, and event sponsorships.
- The item must be about THIS company. If you cannot confirm identity via the domain or unambiguous context, treat it as no news.
- Report the publication date as printed on the source. If a source shows no date, skip it.
- One item only — the most recent that passes the relevance bar.
OUTPUT CONTRACT — follow exactly:
Return a single JSON object:
{"headline": "<as published, under 120 characters>", "date": "<YYYY-MM-DD>", "source": "<publication or site name>", "url": "<direct link>", "category": "<funding | leadership | product | expansion | partnership | award>", "one_liner": "<under 20 words: the fact, stated plainly, usable in an email>"}
- Return ONLY the JSON. No markdown fences, no commentary.
- If no qualifying news from the last 120 days exists or identity cannot be confirmed, return exactly: NOT_FOUND
- Never stretch the window, never report undated items, never summarize the company's own blog as press unless it announces something material.How to use it
- 1
Copy the prompt into Claude, ChatGPT, or any LLM with web access — or into a Clay Claygent column.
- 2
In Clay, map both your company-name and domain columns as inputs — the domain is what powers the identity check, so never run on name alone.
- 3
Spot-check 15 rows: open each URL and confirm the company, the date, and that the one_liner matches the article.
- 4
Merge the one_liner into openers only for rows where category matches your play (funding for expansion pitches, leadership for new-exec plays).
Best practices
Run news enrichment last, days before launch, not when you build the list — a 120-day window ages fast.
Reference the fact, not your congratulations: 'A Series B usually means the SDR team doubles within two quarters' beats 'Congrats on the raise!'
Expect NOT_FOUND on 60–80% of SMB rows; that's healthy. Route them to non-news personalization instead of forcing weak items through.
Example: what this looks like in practice
An SDR at a recruiting firm runs this across 1,800 mid-market accounts on a Thursday, launching Monday. 410 rows return news; 128 are leadership category. She builds a segment for those: 'A new VP of Engineering typically inherits three open reqs and a backfill' merged next to the one_liner fact. One spot-check catches a same-name mismatch the domain check missed for a two-letter company name, so she adds a formula excluding rows where the URL's article doesn't mention the domain. The leadership segment replies at 7.2%, her best cold sequence that quarter, with zero 'that wasn't us' replies.
Best fit
This prompt is one gear in a bigger machine. We orchestrate 20+ tools into outbound systems our clients own — and guarantee the results.
Apply for a Pilot Spot → →Frequently asked questions
Run an AI web agent per row with the company name plus domain, a hard recency window, and a structured output schema. The domain is the critical input — name-only searches misattribute news between same-named companies constantly. In Clay, a Claygent column with this prompt returns headline, date, source, URL, and a merge-ready one-liner.
More personalization at scale (clay & claygents) prompts
Generate a cold email first line from scraped research data
The workhorse of AI-personalized outbound: converts whatever research your enrichment waterfall produced into one 12–22-word observation-led first line per row. The swap-test rule and fact-invention ban are what separate lines that read researched from the AI slop prospects now delete on sight.
Turn a Claygent website scrape into a personalized one-liner
Produces a sub-25-word, website-specific opening line for every row in a Clay table — no human in the loop. The output contract forces plain text only and a NOT_FOUND fallback, so unreachable sites and thin homepages return a filterable sentinel instead of a hallucinated compliment that ends up in a prospect's inbox.
Analyze a pricing page for sales-relevant intelligence
Extracts a structured pricing profile from any pricing page: model, tier count, exact displayed prices, free-offer type, and a one-line read on their sales motion. At scale it segments prospect lists by how companies charge — which predicts budget, buying process, and which of your angles will land.