CX operator + systems builder

Great support is a system. I build it.

I run customer operations, find where they break, and build the software, automation and intelligence systems that make them work better.

See selected work Start a conversation
Running multi-brand ecommerce CX Philippines · GMT+8
support-triage.flow
interactive
Incoming ticket:
Pick a ticket to see where it goes.
AI resolvesroutine & low-risk
Policy checkrules before action
Human nowsensitive or high-risk

    ↑ Scripted demo of how I route support. Pick a ticket.

    • 70%first response time, with AI replying in minutes
    • 100%of tickets triaged by AI: routine ones resolved, high‑risk sent to a person
    • 24hsupport coverage, every day
    • Multi‑brandecommerce CX operations, run day to day

    From my current in-house CX role. Employer and brand names withheld.

    Selected work

    Built because the queue needed it.

    Every project here started as an operational problem I was living with. The problem, what I built, what changed, and what it runs on.

    Internal01 · Case study

    CX Insights / Knowledge Hub

    An operational intelligence layer that turns support tickets into structured knowledge and decisions.

    Problem
    Answers, product feedback and compensation data were scattered across tickets, macros, spreadsheets and people’s heads.
    Built
    AI extracts questions and answers from tickets, a person reviews them, and approved knowledge feeds the team and its reporting.
    Result
    One reviewed source of truth, plus product-question, compensation and voice-of-customer reporting that used to be manual.
    • Zendesk API
    • LLMs
    • Postgres
    • n8n
    • Kubernetes
    Read the case study
    Illustrative view · sample data
    Live02

    AI support automation

    Answers the routine, checks policy on the risky, and gets sensitive tickets to a person fast.

    Problem
    Repetitive tickets (order status, subscription changes) buried the ones that actually needed a person.
    Built
    Classification, policy guardrails and helpdesk actions. AI resolves what’s safe; everything else is routed with context attached.
    Result
    First response time down 70%. Every ticket is triaged by AI: routine ones get a reply in minutes, high-risk ones go straight to a person.
    • Zendesk
    • Shopify
    • Recharge
    • n8n
    • LLMs

    Interface images are illustrations with sample data.

    What I do

    Three layers of a calm support team.

    Good customer experience isn’t just friendly replies. It’s the processes, tools and automations underneath that make those replies fast, consistent and easy to give.

    01

    Customer experience operations

    Workflows, staffing and quality standards that keep response times low and customers happy, across brands and time zones.

    • Support workflows & SLAs
    • Coverage & staffing across time zones
    • QA scorecards & feedback loops
    • Escalation paths & playbooks
    02

    Systems & internal tools

    Connecting the helpdesk, store and subscription stack, and building the internal tools and reporting that sit on top.

    • Zendesk setup & cleanup
    • Helpdesk ↔ Shopify & Recharge data
    • Contact-reason taxonomy
    • Internal apps & dashboards
    03

    Automation & AI

    Workflows and AI agents that triage, resolve and draft, with a human in the loop where it matters.

    • n8n workflows & integrations
    • AI agents with policy guardrails
    • Human review checkpoints
    • Monitoring & safe fallbacks

    How I build

    From messy to maintainable.

    No big-bang rollouts. I start with how the team actually works, then ship small, testable improvements the team can own and run without me.

    1. 01

      Map

      Shadow the queue, read the tickets, and find where time actually leaks.

    2. 02

      Design

      Sketch the simplest process and tool setup that fixes the real problem.

    3. 03

      Build

      Ship in small, reversible steps — with humans checking the parts that matter.

    4. 04

      Hand off

      Docs, training and a clear owner, so the system keeps running without me.

    About

    Operator first. Builder second.

    I run customer operations, then build the systems I wish the team already had.

    My day job is running CX for a group of ecommerce brands: queues, staffing, SLAs, escalations, fulfillment problems, subscriptions, marketplaces, and the reporting on top of it all.

    That’s where the problems worth solving show up. When the fix is a process, I write the playbook. When it’s software, I build it — AI agents, n8n workflows, Shopify apps, Postgres-backed internal tools — and then run it inside the operation that needed it.

    I didn’t learn automation and go looking for problems. I had the problems first.

    How I think

    • Automation handles the repetitive work. Refunds, angry customers and edge cases still reach a person — quickly, with context attached.

    • The best workflow is the one your team understands and trusts. Simple, visible and easy to switch off if something goes wrong.

    • Every process and automation ships with a short, plain-language doc so your team can run it, change it and fix it.

    • Fast replies are great. Fewer repeat contacts and happier agents are better. I track both.

    What I build with

    Tools matter less than knowing where the operation breaks. These are the ones I reach for.

    CX 01

    • Zendesk
    • Help center & macros
    • Triggers & SLAs
    • QA scorecards

    Commerce 02

    • Shopify
    • Recharge
    • Klaviyo
    • Marketplaces
    • 3PL & fulfillment

    Automation & AI 03

    • n8n
    • LLMs
    • Classification & extraction
    • Webhooks & APIs
    • Human-in-the-loop review

    Engineering 04

    • Postgres & SQL
    • Kubernetes
    • Internal web apps
    • Shopify apps
    • GitHub Actions

    Contact

    Got a similar mess?

    I’m always up for comparing notes. Talk to me about:

    • CX systems
    • AI automation
    • Internal tooling
    • Ecommerce operations
    hello@vincentalcordo.com
    Philippines — GMT+8 It’s --:-- here right now