Key takeaways
- Most manual effort goes into research and first drafts, not the final edit.
- The problem, the proof and the structure can be written once per segment.
- People still need to choose accounts, judge triggers and read the final draft.
- If replies show prospects do not know why you wrote, you have cut too much.
Where does the manual effort in personalisation go
Most of the time goes into research and first drafts, not into the final edit. A seller opens the prospect's website, scans their LinkedIn, looks for news, then faces a blank email.
Each step is small, but repeated for every prospect it fills the day and leaves little time for conversations.
Which parts can be done once per segment
The problem, the proof and the structure can be written once per segment and reused. A message to finance leads at mid-sized software companies shares most of its content with every other message to that group.
Write that version carefully, with a real example, and it carries most of the relevance before any personal detail is added.
Which parts still need a person
A person should still choose the accounts, judge the trigger and read the final draft. AI can find and summarise information, but it cannot reliably tell whether a trigger is a good reason to write.
A funding announcement from two years ago, a post the prospect shared without comment or a job advert already filled all look like signals and are not.
How do you know when you have cut too much
You have cut too much when replies show the prospect did not see why you wrote to them. Watch for replies asking what you do, unsubscribes from people who clearly fit, and drafts you would be embarrassed to see forwarded.
Any of these means the personal line is missing or fake, so add research time back to the top tier first.
How to personalise with less manual effort
Set up the first three steps once; repeat the last two for each batch.
- Tier your accounts by fit, and decide which tier gets hand-written messages.
- Write one strong message per segment, with a real example and a clear ask.
- Agree what counts as a usable trigger, and what does not.
- Gather triggers and company research for a whole batch at once, then have AI draft the personal line.
- Review each draft, fix or delete the personal line, and send from your own mailbox.
Where Pebbles Ai fits
- CapabilityProspecting & customer recordsDescribe your ideal customer; your assistant finds matching companies, then the right people at up to ten of them. Each lead arrives with personalisation angles from their public LinkedIn posts, or from their role and company.
- FoundationCompany IntelCompany Intel holds your Business knowledge and Brand voice, so you stop re-explaining your company in every request.
- CapabilityOutreach contentCampaign drafts follow the personality you choose for each piece and are saved as drafts in your connected Gmail or Outlook mailbox, ready for a quick edit.
Common questions
Can AI personalise outreach without any human review?
It can produce drafts, but a quick review catches stale triggers, wrong facts and lines that sound off. Keep a person on the send decision, at least for your best accounts.
How much research does one prospect need?
Enough to find one real reason to write. If nothing turns up after a short look, send the segment message instead.
Is segment-level personalisation really personalisation?
Yes, when the segment is narrow. A message written for one role in one industry with one problem is more relevant than a generic email with the prospect's name in it.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .
