The Setup
Outbound sales for a staffing and AI company means reaching engineering leaders at the right moment — when they're hiring, scaling, or dealing with a workflow problem. Getting the timing and message right manually is slow. Getting it wrong with automation is worse.
The Problem With AI Email Tools
I've tried most of them. They all work the same way: import a CSV, pick a template, insert {{first_name}} and {{company_name}}, hit send. Sometimes they'll add a "personalized" line pulled from LinkedIn. It still reads like a mass email — because it is.
The email that gets a reply isn't the one that says your name. It's the one that says something true about your company that the sender clearly spent time finding out. That's what we wanted to automate — not the sending, but the actual research.
What We Built
For each lead in our list, the sales outreach agent does real company research before writing a single word:
- →Reads the company website to understand what they do and who they serve
- →Pulls open engineering job postings to identify what roles they're actively hiring for
- →Checks for recent funding announcements or growth signals in public sources
- →Looks at the engineering blog or technical posts if they exist
- →Uses all of that to write a 4-sentence email in our voice — specific to that company, specific to that contact
The result gets loaded as a draft. I review it, make small edits if needed, and send. The agent does the research and the writing. I do the final judgment call.
The 4-Sentence Rule
Every email the agent writes follows the same structure: one sentence showing we know something specific about them, one sentence about the problem we solve, one sentence on why we're different, and one sentence with a clear ask. Four sentences. Nothing more.
The constraint is intentional. Long emails don't get read. The agent can't ramble because the template doesn't give it room to ramble — and the constraint forces it to choose the most relevant signal from everything it researched.
What Changed
- →Time per researched draft: from 15–20 minutes to seconds
- →Volume: 50 researched, ready-to-review drafts in under 10 minutes
- →Quality: better than what I was writing manually, because the agent doesn't get lazy on draft 40
- →Outreach volume: scaled significantly without adding anyone to the team
What We Learned
The research quality matters more than the writing quality. When the agent finds a weak signal — a company with no blog, no job postings, no recent news — the email is weaker. We added a quality filter so those leads get flagged for manual review instead of auto-drafted.
The agent also needed to understand what we do well enough to match signals to offerings. "They're hiring 3 backend engineers" → mention offshore engineering. "They just raised a Series A" → mention speed of placement. Teaching the agent to make those connections was the real work.
Want This for Your Sales Outreach?
If you're doing any kind of outbound — sales, partnerships, recruiting — we can build a version of this for your workflow. Book a scoping call and let's talk specifics.
Book a Scoping Call →