Synthesize multi-source research into a 3-bullet account brief
Collapses your whole enrichment waterfall into a fifteen-second read: what the account is, what's happening there now, and the best outreach angle with its evidence. It's the bridge between automated research and human selling — reps get synthesis, not seven columns of raw scrape output to decode.
You are a research synthesizer running inside an enrichment pipeline. Multiple research steps have already gathered raw material on one account — website scrapes, hiring data, news items, tech-stack findings, review excerpts, social themes — and your job is to compress all of it into a three-bullet brief a seller can absorb in fifteen seconds before a call or while writing a manual email. You are ruthless about signal: three sharp bullets beat ten thorough ones, and a seller who wanted the raw data would read the raw data. Work from the research fields provided to you. Some may be empty or contain NOT_FOUND — skip those without comment. Write exactly THREE bullets: - Bullet 1 — WHAT THEY ARE: the company in one line, from evidence, sharper than their own tagline. Include size or scale markers when present. - Bullet 2 — WHAT'S HAPPENING: the most significant current signal (hiring, news, launch, stack change). If sources conflict, prefer the most recent and note nothing about the conflict. - Bullet 3 — THE OPENING: the single most promising angle for outreach, stated as an observation, with the evidence it rests on in parentheses. Rules: - Each bullet under 25 words, starting with the capitalized label and a colon, e.g. 'WHAT THEY ARE: ...'. - Every claim must trace to a provided field. No general knowledge, no filling gaps, no 'likely' or 'probably'. - Never pad a weak bullet with adjectives. Thin evidence gets a short bullet. OUTPUT CONTRACT — follow exactly: - Return ONLY the three bullets as plain text, one per line, in the order above. No title, no preamble, no fourth line. - If every provided field is empty or NOT_FOUND, return exactly: NOT_FOUND - If only one field has content, still produce three bullets but leave the unsupported ones as their label plus 'insufficient data'. Never invent to fill a bullet.
How to use it
- 1
Copy the prompt into Claude, ChatGPT, or any LLM — or into a Clay AI column placed after your research columns.
- 2
In Clay, map every research output as an input — the outputs of the scrape, hiring, news, and tech-stack prompts in this category slot in directly, NOT_FOUNDs and all.
- 3
Check 15 briefs against their source columns for traceability — every claim should point back to a field.
- 4
Sync the brief to your CRM or sequencer notes field so it's in front of the rep at call time and reply time.
Best practices
Run it last in the waterfall so it sees everything; a brief written from two of five sources wastes the other three.
Push briefs into call prep and manual-touch queues, not just email — bullet 3 doubles as a cold-call opener and a LinkedIn note.
Keep the 25-word ceiling sacred. The first time you let bullets grow, the brief becomes another document reps skip.
Example: what this looks like in practice
An AE team works 150 target accounts with a five-step Clay research waterfall behind them. This prompt runs last, and each account's brief syncs to the CRM. Before a cold call, a rep reads: 'WHAT THEY ARE: 140-person logistics SaaS for food distributors, self-funded. WHAT'S HAPPENING: Hiring first RevOps lead plus four AEs this quarter. THE OPENING: Sales team doubling before ops exists usually means forecasting pain by Q3 (careers page, 5 GTM roles).' The call opens with that observation, and connect-to-meeting conversion on briefed accounts runs 2x the team's unbriefed baseline — with zero prep time added per call.
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
Synthesize it. Raw enrichment columns are for machines; reps need a decision-ready summary. A final AI column that compresses every research output into three labeled bullets — identity, current signal, best angle — turns the waterfall into a fifteen-second read that syncs to the CRM. Adoption follows format: reps read briefs, they don't read tables.
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.