ABBIE ABANES
Ops Automation

Live System · Solo Build

Every request logged.
Every deadline watched.
Nothing manual.

A small ops team was reading every client request by hand, replying by hand, and only finding out about a missed deadline when a client complained. I built a simple system that takes a request from form submission to a drafted reply and a watched deadline - no manual step in between.

Client Ops Automation AI-Assisted Drafting Built & shipped solo Live

01 · The Problem

Nothing was watching the clock

Every request got the same manual treatment - read it, reply to it, move on - no matter how urgent it actually was. A deadline only became visible once it had already passed, usually because a client asked what happened.

02 · What I Built

One pipeline, one clock, no manual handoffs

A request moves from form to logged record to a drafted reply that's one tap away from actually going out - and a second, separate check watches for anything that's gone overdue and flags it on its own.

On every new request
Tally
Client submits
→
n8n
Parses & sets deadline
→
Airtable
Logs the record
→
Gemini
Drafts a reply
→
Slack
Draft + Approve button
→
Email sent
Goes to the client on tap
Watching for overdue requests
Schedule
Hourly / daily
→
Airtable
Finds overdue items
→
Airtable
Marks escalated
→
Slack
Alerts the team
Estimated time per request
Manual: ~12 min per request. With this system: ~2 min. That's roughly 10 minutes back per request - across a queue of a hundred requests, that's over 16 hours returned to the team.
01

AI drafts, a human sends

Every new request lands in Slack with the client, priority, deadline, and a ready-to-edit AI draft - a reply is one read-and-send away.

Slack message in #new-requests showing client, type, priority, SLA deadline, description, and AI-drafted response
A real request, logged and drafted automatically.
02

One table, every request

Every request lives in a single Airtable with its own deadline and status - New, In Review, Escalated, Closed.

Airtable Client Requests table showing client name, request type, priority, description, submitted date, SLA deadline, status, and AI draft response
Live data from the system in use.
03

Overdue requests flag themselves

Checks run hourly for urgent requests and once a day for standard ones (configurable) - anything overdue gets marked and posted straight to the team.

Slack SLA breach alert in #escalations with client, priority, request type, and the missed deadline
A real overdue request, caught without anyone checking manually.

03 · Impact

What changed once it shipped

~10 min
Saved per request
0
Missed SLAs since launch
100%
Requests logged automatically
1 tap
From drafted reply to sent

04 · Title Block

Built plainly, on purpose

Simple, reliable tools - easy for a small team to maintain after handoff.

Intake formTally
AutomationRuns continuously in the backgroundLogs every request before drafting a reply, so a slow AI call never blocks a request from being tracked
Record keepingAirtable - one table, every request and its status
AI draftingGemini drafts the first reply - a human always reviews before sending
Approval & sendSlack Approve button triggers the actual email to the clientOne tap replaces copying the draft into an inbox by hand
Escalation cadence1 hour for urgent requests, once daily for standard ones (configurable)

05 · Get In Touch

● Live

Let's talk about what's
broken in your process.

Whether that means a new system or just a better way of working, I'd rather start with what's actually broken than pitch you a build you don't need.

Abbie Abanes
work.abbiea@gmail.com
+63 995 524 3072
Mo—Fr 9am—6pm PHT
Abbie Abanes.