Completed workflow (Built)

AI Cold-Outreach Draft Helper

Writing cold emails that don’t sound like cold emails takes time. The research and first draft are the slow part, and that’s exactly what can be helped.

Status
Completed workflow
Area
Sales & Outreach Automation
Workflow
5 steps · 3 failure paths
Stack

01 · Why it mattered

What the manual way was costing

  • Personalised outreach didn’t scale past a few per day.
  • Template emails were fast but obviously templates.

02 · System concept

What I set out to build

Go through the sheet row by row, qualify the prospect, write a personalised draft, put it back in the sheet, and tell me if anything breaks.

03 · Before / after

From manual steps to a system

Toggle between the manual process and the system. Click any step to see what it does, what tools it touches, and what happens if it goes wrong.

How it ran manually · depends on whoever is free

  1. 01Prospect list sits in Google Sheetsby hand
  2. 02Each prospect is researched manuallyby hand
  3. 03Each email is written from scratch or from a stiff templateby hand
  4. 04Nothing records which rows are doneby hand

04 · Decision log

Why I built it this way

D-01Drafts, not auto-send

I deliberately stopped at the draft. Outreach goes out under someone’s name, so a person should read it first.

D-02Qualify before writing

No point writing a lovely email to the wrong person. Qualification runs first so drafts only get written for rows worth it.

D-03Errors should be loud

A workflow that fails silently is worse than no workflow, because you think it ran. I added error notifications for that reason.

05 · Reliability layer

What can go wrong, and who steps in

Where it happens
Step 02 Qualification
What goes wrong
Prospect doesn’t meet criteria
What the system does
Row is marked as not qualified and no draft is written.
Who handles it
The system handles it on its own

Human-in-the-loop points

  • Human in the loopHuman reviews & sends
  • Human in the loopRow is missing data needed to personalise → me
  • Human in the loopAI or API call fails → me

06 · Evidence

What exists today

  • Completed workflow, running from a Google Sheet with error notifications.

07 · What this demonstrates

What this demonstrates

Using AI for the slow part of a task while keeping the human decision exactly where it matters.

08 · Stack / integrations

What it’s built with

  • Google SheetsProspect list and draft output
  • OpenAIQualification and draft writing
  • GmailSending (after review) and error notifications
  • n8nOrchestration

Architecture · static view · 5 steps · 3 failure paths

    • → meRow is missing data needed to personalise
    • → system handlesProspect doesn’t meet criteria
    • → meAI or API call fails

More systems

Keep reading

Client workflow / pilot

Dental Patient Reactivation System

Patients who are overdue for a visit get a personal follow-up with a booking link, and the clinic can see who booked from one Airtable view.

  1. Identify overdue patients
  2. Personalised follow-up
  3. Booking link
  4. Booking tracked
  5. Clinic sees results
3 failure paths · 2 human handoffs
Deployed application

RecruitTech ATS

A deployed applicant tracking app where every application becomes a candidate record moving through screening and interview stages.

  1. Application
  2. Candidate record
  3. Screening / status
  4. Interview stage
  5. Recruiter action
  6. +1 more
3 failure paths · 2 human handoffs

Similar workflow?

Got a process that looks something like this?

Tell me how it runs today and where it hurts. If it’s a good fit for automation I’ll tell you how I’d approach it. If it isn’t, I’ll tell you that too.

Tell me what you’re trying to automate