Researched, not merge-tagged.
Every BongoBot email is written using real data from the prospect's website — their services, their target market, what makes them different. The result reads like you spent ten minutes researching each lead, not like a template with {first_name} swapped in.
An email, with sources
Below is an outreach email written by BongoBot to a Bristol plumbing company. Every highlighted phrase traces back to a specific field extracted from their website. Toggle to “Mail merge” to see what a template-based tool produces instead.
Noticed Hill & Sons has been doing bathroom refits around Bristol — and that your family has been at it since 1987. That kind of track record is rare.
The problem: most Bristol homeowners searching for emergency plumbing never see family-run trades — the chains own page one. We help shops like yours rank for local search without a contract.
Worth a quick chat?
How it works│When a prospect is discovered, BongoBot scrapes their website and runs it through an AI extraction step that produces structured data: products/services, target market, differentiator, location. That structured data is passed directly to the email generation prompt as context — so the writer has real facts to reference, not just tokens to swap in.
From URL to email, in four steps
Every prospect goes through the same flow before a single word is drafted. The writer only ever sees structured, verified facts — not a raw HTML dump.
Scrape
BongoBot fetches the homepage, about page, services page, and any landing pages it discovers. Raw text only — no screenshots, no tracking.
Extract
A Claude call parses the scraped content into four structured fields: products/services, target market, differentiator, and a 1-2 sentence summary.
Write
The structured data is passed to the email generation prompt as factual context. The writer references real details instead of relying on merge tags.
Fact-check
A second AI pass checks the draft against the extracted facts. Invented stats, wrong numbers, or off-topic claims get flagged and rewritten.
Why BongoBot emails don't sound like BongoBot emails
The generation prompt has strict constraints that rule out the phrases and structures that tank open rates.
- • 60-75 words total, 80 hard max
- • Subject line 3-6 words
- • 2-3 short paragraphs after the greeting
- • First sentence references the prospect, not the sender
- • Must use both {lead_name} and {contact_name}
- • Plain text, no formatting, no emoji
- • “I hope this finds you well”
- • “unlock”, “transform”, “revolutionize”, “skyrocket”
- • “In today's [competitive] landscape”
- • “Book a demo” / “Schedule a call” / “Limited spots”
- • Invented percentages or ROI claims
- • Claims about what “most” businesses do
Fact-check pass
Before the email gets queued, a second AI pass reviews it against the extracted facts. If the writer invented a statistic, claimed something about the prospect that isn't on their website, or slipped into a non-English market with English copy, the draft gets rewritten. Honest claims only.
Your first campaign
starts in minutes
Every week you wait is another week your competitors are in their inbox. 100 real prospects that match your ideal customer — free.