PM.← All work

AI-assisted support · ChatDaddy

From Architect Suite to the production Backend Doctor

Originated an investigation system adopted into a production support tool that correlates Kibana, frontend, and backend evidence.

diagnosis, down from 1–2 hours10–15 min

01 · Context

What needed to change

Support and engineering teams investigated recurring messaging issues across many services while a backlog of 30–40 tickets persisted.

02 · Challenge

The hard part

Identify root causes across logs and services, then prepare a useful customer response while keeping a person in the approval loop.

03 · Approach

How I approached it

  1. Created Architect Suite to correlate frontend, backend, and Kibana evidence across 22 services and 257 patterns.
  2. Produced investigation notes covering symptoms, likely cause, and customer-facing response.
  3. Used a human approval step before the support workflow advanced; the company adopted the code into its production Backend Doctor.

04 · Outcome

What shipped

AI-assisted diagnosis fell from 1–2 hours to 10–15 minutes, and the ticket backlog fell from 30–40 to 2–3.