Scaling an IT services firm's personalised outreach with a drafting agent and human approval
Research-led messages earned replies, but each needed a skilled writer. A drafting agent now writes from the brief in the firm's voice, and nothing sends without approval.
- Challenge
Personalised outreach worked, but every message needed a skilled writer's time.
- Approach
A drafting agent configured on the firm's voice, with approval before any send.
- Result
Drafts arrive ready for review; SDRs approve, edit or reject each one.
The situation
An IT services firm's best outreach is written by hand from a research brief: a trigger, a named buyer and a relevant delivery story. Replies are good, but each message takes a skilled person's time, so volume grows only by hiring. New SDRs need time to learn the firm's voice and which proof points suit a CIO, a CTO or a head of digital. Generic templates would scale, but they are exactly what enterprise IT buyers already ignore.
The question that mattered
Can drafting scale without diluting the voice and research that make the outreach work?
What we did
- 1
Capture the voice
Build a voice and claims guide from the firm's best-performing messages, covering tone, structure, approved proof points and phrases to avoid.
- 2
Build the drafting agent
Configure an agent to turn each research brief into email and LinkedIn messages, drawing only on the brief and the approved proof library.
- 3
Put approval before sending
Route every draft to an SDR approval queue with no auto-send. Drafts for senior executives or existing customers escalate to an account lead.
- 4
Learn from edits
Log each draft against the version actually sent, review recurring edits, and update the voice guide and prompts under supervision.
What we delivered
The evidence and working documents the team kept.
- Voice and claims guide built from the firm's own best messages
- Approved proof library mapped to sectors and buyer roles
- Drafting agent for first-touch and follow-up messages on email and LinkedIn
- Approval queue with escalation rules and no auto-send
- Draft-versus-sent log used to refine the guide and prompts
The outcome
The agent
drafts each message from the research brief in the firm's voice
SDRs
approve, edit or reject every draft before it is sent
Missing brief details
are flagged for the researcher, not invented
Messages to senior executives
need an account lead's sign-off
Outreach still reads as the firm's own, and SDR time moves from writing to judging fit and timing.
What changed in the GTM motion
Personalised outreach is no longer limited by how many people can write it. The voice guide, proof library and edit log become shared assets that new SDRs learn from, and the agent changes only through reviewed updates.
Services in this programme
Case studies
More programmes like this
Outreach that works but only scales with headcount?
Tell us where it is stuck. A founder replies within one working day.