Candidate Outreach Automation for Recruiters: Research, Personalise, Follow Up, Stop on Reply

A recruiter can write ten good outreach messages a day or fifty bad ones. The good ones get replies; there is never time for them. The follow-ups, where most positive replies actually arrive, are the first thing dropped when the week fills up. And the ATS shows "contacted" next to people who got one message six weeks ago.

Outreach automation is usually sold as a way to send more. The version that works is a way to send better at the same volume: a researched first message, two follow-ups that actually happen, and a hard stop the moment the candidate answers. This post covers how the candidate outreach blueprint does that and where the limits are.

One specific fact is the whole difference

Candidates can tell within one line whether a message was written for them. "Saw the migration you led at Northwind" gets a reply; "I came across your profile" does not.

So the automation spends its effort on research: the public profile, the company page, recent talks or posts, open-source or portfolio work. From three to five facts it picks one hook: something specific, recent and professional that relates to the role. Not the title, not the employer's tagline.

Then it refuses to send when it cannot find one. A generic message costs more in reputation than it earns, and the recruiter would rather see "no hook found, skipped" than a template going out under their name.

The sequence

1. Receive the candidate from a sourcing project in Greenhouse, Lever or Ashby, or a row in a sheet.
2. Research public sources and summarise three to five facts, with links.
3. Pick the hook.
4. Draft under 90 words: the hook, why the role is relevant to it, one question. Plain text, no tracking pixels, no "I hope this finds you well".
5. Send from the recruiter's own address, spaced through working hours.
6. Follow up twice, day 4 and day 10, each from a different angle (the team, the problem, the timing), each shorter than the last.
7. Stop on reply. Any reply ends the sequence and forwards the thread with the research to the recruiter; an out-of-office pauses it.
8. Measure per role: sent, replied, positive, interviews, by hook type.

The last step is what makes the second month better than the first. When you can see that "recent talk" hooks outperform "current project" hooks for engineering roles, you write the criteria for the next sourcing list differently.

Sending limits protect the address you cannot replace

The recruiter's own address is the asset. Forty new messages a day, spaced, plain text, no tracking, from a mailbox that already sends normal mail is what keeps that address landing in inboxes. The automation enforces the limit rather than trusting the day's enthusiasm, and every message carries an opt-out line that, when used, suppresses the candidate across every role.

LinkedIn is deliberately not a channel in the standard version. It has its own limits and tooling, and a blocked LinkedIn account is worse than a slower start. It is a reasonable custom addition once the email version has run for a month.

What "not automated" means to the candidate

The candidate sees a message from a named person, about something specific they did, that stops the moment they write back. From the reply onwards it is the recruiter, with the research already in front of them. Candidates notice generic messages, not well-researched ones, and they notice being chased after they said no; the stop-on-reply rule handles both.

Where it fits with screening and scheduling

Outreach is the top of the pipeline. The resume screening blueprint scores the applications that come back against the role's criteria, and interview scheduling automation books the conversations. Together they are the recruiter package: three blueprints, installed in your ATS in about a week.

For one recruiter the outreach blueprint alone runs as a managed automation at $197 a month, or as a kit and install for people who would rather own it. An agency with several recruiters sharing sequences, outreach in several languages, or LinkedIn steps are a custom build.

See also AI HR agents.

Sources and further reading

• Greenhouse, Harvest API — the write-back endpoints the blueprint uses for notes and candidates: https://developers.greenhouse.io/harvest.html
• New York City Department of Consumer and Worker Protection, Automated employment decision tools (Local Law 144) — relevant once outreach feeds an automated screen: https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page