SDR First Touch Brief - letters

The problem

Cold outreach that opens with "I noticed your company is doing great things" gets deleted in two seconds, and manually researching a prospect's site, news, funding, and hiring pages for a genuinely personal hook eats 15-20 minutes per account — time most reps don't have across a real prospecting list.

What it does

• Fetches the target company's own homepage, about, news, customers, and careers pages, with WebSearch fallbacks when a page 404s.
• Extracts every distinct qualifying item against a fixed 10-category checklist — recent news, product specifics, leadership, funding, hiring, customers, a differentiator/process signal that adapts to the target's own domain, headcount, positioning, and (if a prospect name is given) a corroborated prospect-specific signal — not just one item per category, so a company with three separate newsworthy items gets all three considered.
• Labels every extracted item — not just the ones it ends up using — as relevant or not relevant to your stated product, with a one-clause reason, shown in a Research Log audit table; when a product is given, only relevant items are eligible for the drafted emails.
• Selects the 3 most specific, relevant facts and drafts a fixed 3-email sequence: a trigger-based opener (~70-100 words), a value-add follow-up (~60-90 words), and a low-pressure breakup (up to 100 words) — the 3 facts can come from a single category (three separate news items, say) or be spread across several; category diversity is never required, only genuine relevance and distinctness.
• Runs a diversity and generic-filler check so no two emails reuse the same fact and no sentence matches known cold-email boilerplate.

Why this beats prompting it yourself

A naive "write me a cold email for {company}" prompt fills gaps with generic flattery the moment research comes up short, defaults to tech-company language regardless of what the target business actually does, and gives no way to check why it picked what it picked — or arbitrarily forces itself to spread facts across categories even when the strongest signal is a cluster in one. This skill instead requires at least 3 real, sourced, product-relevant facts before it will draft anything, labels every signal it found — used or not — as relevant or not in a visible audit table, and lets the strongest facts win regardless of which category they came from.

Use cases

• An SDR building a first-touch sequence for a new named-account list spanning multiple industries.
• A founder doing their own outbound before hiring a sales team.
• An agency drafting outreach on behalf of a client with a specific product angle.
• Refreshing a stale sequence template with facts pulled from a prospect's latest funding round, product launch, or a cluster of recent news.