ERPNext for a Small Team
An AI assistant for everyday ERP work, from approvals to inventory decisions.
“Some stock hasn’t moved in a while. Should we buy more or hold off?”
Ask in everyday language
The assistant checks the records you have permission to see, explains what it finds and suggests what to review next.
1 Ask
Ask which items are selling slowly.
2 Review records
Compare stock on hand with sales over the same period.
3 Decide
Use the evidence to reduce or postpone the next order.
Slow-moving stock
A stock count tells you what is on the shelf. Comparing it with sales shows which items tie up cash while selling slowly. The assistant lists those items and the figures behind its suggestions so the manager can decide what to reorder.
Approval delays
Ask which approval steps have the longest waits. The assistant uses workflow records to flag bottlenecks so the team can investigate the cause and adjust how requests are handled.
Business challenge
This business of fewer than ten people needed one system for HR, finance, stock and sales. A fully custom ERP exceeded its budget. Standard SaaS products did not fit several approval processes, and staff had little time to learn new menus and forms.
Solution
After mapping the team’s workflows, I deployed ERPNext and an AI agent on an Ubuntu mini server. I tailored HR, finance, inventory, sales reporting and customer management, added an internal knowledge base and connected Lark for single sign-on and synchronized role permissions.
Daily use
Employees describe a task to fill in and submit forms. Authorized colleagues can review and approve them in conversation. OCR extracts receipt details to reduce manual entry. A request such as “Rank customers by total sales this month” returns a report within the employee’s access rights.
Project facts
Fewer than 10 staff
1 month to deliver
US$800 total
3 months of support