[03]AI · Automation · Operations
30+ bots and AI agents in production
Turning a constant stream of ERP transactions into fast decisions instead of repetitive work.
- Organization
- YoLa Fresh
- Role
- Tech Manager — project lead
- Period
- Nov 2024 — Present
30+
bots and AI agents
50,000+
ERP transactions a day
9
sales regions
< 60 s
to an actionable alert
01The challenge
How do you absorb more than 50,000 ERP transactions a day, across 9 sales regions, without reconciliation, reporting and stock tracking turning into manual work?
02My role
- 01Built and industrialized more than 30 bots and AI agents
- 02Automated high-volume repetitive work
- 03Real-time alerts in Google Chat, with owner and context
03The system
From signal to action
- 01
Business problem
Repetitive work
- 02
ERP data
50,000+ tx / day
- 03
Bot / agent
30+ in production
- 04
Decision
Owner + context
- 05
Business action
< 60 s
- Cash reconciliation
- Daily P&L
- Stock alerts
- KPI reporting
Every automation follows the same path: a business problem, ERP data, a bot or agent, a decision, an action.
04The execution
Technology
- Node.js
- Python
- SAP Business One
- Cron jobs
- Google Chat
People
Every alert arrives with its owner and context: the team acts instead of searching.
Process
Concrete use cases — cash reconciliation, daily P&L, stock alerts, KPI reporting — industrialized one by one.
05The result
- More than 30 bots and AI agents in production
- More than 50,000 ERP transactions processed a day across 9 sales regions
- Actionable notifications in under 60 seconds