- The 60/30/10 rule for outreach personalisation
- How to use AI as a draft, not a send button
- A subject line formula that lifts open rates by 40%
- Five signals that beat 'I came across your profile'
You already know what bad outreach looks like. “I came across your profile and was impressed by your background.” Delete. “Quick question for you.” Delete. “Are you open to new opportunities?” Delete.
Candidates have been trained to ignore those messages. AI did not invent the problem. AI just makes it possible to send 100 of them in the time it used to take to send 10.
The fix is not to write fewer messages. The fix is to make every message feel like it was written by a human who actually read the profile. Here is the system.
The 60/30/10 rule
Every outreach message should be roughly:
- 60% about them. Something specific you noticed. A project they shipped. A talk they gave. A company they joined.
- 30% about the role. Why this fits their trajectory, not why this is a great company.
- 10% about the ask. Make the next step tiny. 'Worth a 15-minute chat?' beats 'Let me know if you'd like to learn more.'
How to actually use AI here
The mistake is asking AI to write the whole message. The right move is to ask AI to find the personalisation hook, then write the message yourself.
Good prompt: Read this candidate’s profile and tell me three things they have done that I could reference in an outreach message. For each one, explain why it matters for the role I am hiring for.
Now you have three real hooks. Pick the strongest, write 90 words around it, and send. Total time: under three minutes per candidate. And the message will land.
The subject line formula
Open rates live and die on the subject line. The formula that consistently lifts opens by 30 to 40%:
[their name] - [one specific thing about them]
Examples:
- Léa - your Sigma migration post
- Tariq - the Adyen → fintech path
- Mei - shipped fast at N26
Avoid: “Quick question”, “Opportunity at [company]”, “Are you open?”. Every recruiter sends those. Yours should not.
Five signals that beat ‘I came across your profile’
The opening line decides whether they read the rest. These five signals consistently outperform generic openers because they prove you read the profile.
- Reference a specific project they shipped, not just a job title.
- Mention a company they were at and what that company is known for, not just the name.
- Reference a post they wrote or a talk they gave on a topic relevant to the role.
- Note a transition pattern (e.g. early-stage to scale-up) and tie the role to that arc.
- Acknowledge something they care about (open source, mentoring, a specific tech) before you mention the role.
Multi-channel sequencing
One channel is rarely enough. The cadence that works for most senior candidates:
- Day 1: email with the strongest hook + soft ask
- Day 4: LinkedIn connect with note (different angle, not a copy of the email)
- Day 9: short follow-up on whichever channel they engaged with
- Day 16: final value-add (an article, a market signal, a relevant data point) with no ask
Stop after four touches if there is no engagement. Move on. Re-engage in 6 months when their context might have changed.
What ‘at scale’ actually looks like
A solo recruiter using this system can run 30 to 50 personalised first-touches a day. AI does the profile reading and signal extraction. The recruiter spends 90 seconds per message picking the right hook and tightening the wording. The result feels handwritten because the most important part of it is.
The bottom line
AI does not write better outreach than you. AI gives you the ammunition to write outreach faster than you ever could before. The recruiter is still the brain. The AI is the research assistant.
