Background

Outbound is a numbers game with a cruel catch: the work that actually moves the needle — researching the right companies, finding the right person, writing something they'll reply to, and then chasing every response — is exactly the work nobody has time to do at volume. So most outbound ends up either low-volume-and-personal or high-volume-and-ignored.

I wanted to collapse that tradeoff for my own pipeline (contract work and roles): a system that runs outbound the way a disciplined SDR would — research each target, personalize every message, send from a real inbox, then handle whatever comes back — except it never sleeps, never skips a follow-up, and costs effectively nothing to run.

That system is Rainmaker. It prospects, personalizes, sends, and then reads and triages every reply — fully automated, end-to-end, on serverless infrastructure I own.

Challenge

The reason "just automate your outreach" isn't a solved problem — and why most attempts turn into spam within a week — comes down to a handful of walls:

• Volume kills personalization, normally. Blasting a template gets you spam-foldered and blocked. Personalizing by hand doesn't scale. You need genuine per-target research done automatically.
• Finding the right contact is half the battle. A company name isn't an email address. Getting to the actual decision-maker's inbox is its own pipeline.
• The inbox is where automation usually stops. Sending is easy; understanding the replies — "is this an interview invite, a rejection, a real lead, or noise?" — and reacting correctly is the hard, human-feeling part.
• Deliverability punishes bots. Send like a machine — same cadence, mechanical fingerprints — and you land in spam or get your domain torched.
• It has to run unattended, forever. A script you babysit isn't automation. A cron job on your laptop dies the moment you close the lid.

Solving any one of these is a weekend project. Solving all five at once — and keeping it running unattended — is the actual work.

Solution

Rainmaker treats outbound like a data pipeline: research in, personalized touches out, replies classified and routed.

Target research, automatically

For each company in the CRM, it fingerprints the tech stack — cloud provider, infrastructure-as-code, backend language, key services — so outreach can be targeted and personalized to what the company actually runs, instead of a generic blast.

From a company name to the right inbox

It resolves each target to a real contact via Apollo.io, then cleans and normalizes the messy inputs — company names, job titles, staffing-agency noise — and filters against blacklists so nothing goes to the wrong place.

Sends from real inboxes, like a person

Personalized emails go out through actual Gmail inboxes — several of them — paced and shaped to read like a human wrote them, not fired from a template cannon.

Reads and triages every reply with an LLM

This is the part most tools skip. A serverless classifier (AWS Lambda, on a 15-minute schedule) polls the inboxes and uses Claude to read each inbound message and decide what it is — interview invite, rejection, real lead, LinkedIn InMail, or noise — then routes it accordingly: logs interviews, tracks rejections, drafts the right reply. It even parses the gnarly HTML of LinkedIn and job-board notification emails to pull out the real content underneath.

Built to run itself

The whole thing lives on serverless AWS — Lambda on a 15-minute EventBridge schedule, an SQS queue for backfill, DynamoDB for idempotent dedup so nothing is ever processed twice — all defined in Terraform. It survives restarts, handles concurrency, and needs zero babysitting.

Results

Rainmaker runs my outbound end-to-end with no one at the keyboard: it researches targets, finds the right contacts, sends personalized outreach from real inboxes, and then reads and triages every reply — around the clock, at effectively zero ongoing cost.

Why this matters for a business:

• Outbound is a throughput problem, not an ideas problem. The bottleneck is never "who could we reach" — it's the human hours to research, personalize, send, and follow up. Automate that and the ceiling moves.
• The reply-handling is the moat. Anyone can send email. A system that understands the responses and reacts correctly is what turns sends into conversations.
• The economics invert. When a researched, personalized touch costs ~$0 and zero human minutes, "reach more people" stops being a headcount decision and becomes a switch you flip.

Conclusion

Rainmaker is proof that outbound — the part everyone insists "can't really be automated without turning into spam" — can run itself when you treat it as an engineering problem instead of a copy-paste chore. I built it to run my own pipeline; the same engine is an always-on outbound/SDR system for any business that wants a steady flow of conversations without hiring a team to grind them out.

If you want an outbound engine like this pointed at your market — researched targets, personalized sends, and AI that handles the replies — that's exactly what I do. Book a call and I'll walk you through what a pipeline like this could do for your funnel.