Personal project
Every order in one place and one format.
Orders arrive from the shop, the mailbox and Instagram messages. The system normalises them into a single format, detects duplicates and lines them up in one fulfilment queue.
The problem it solves
When orders come through three channels somebody has to retype them into one spreadsheet. That takes time and creates errors: a wrong quantity, an order lost in a direct message, the same customer served twice.
Built with
- n8n
- Google Sheets
- Shop APIs
- Slack
Related service
Automation and AI agentsHow it works
- 1Each channel has its own reader: the shop via API, email via the mailbox, messages via an integration
- 2The data is normalised into a shared format with the same field names
- 3The system compares new orders against existing ones and flags possible duplicates
- 4Complete orders join the fulfilment queue; incomplete ones wait for missing details
- 5A status change sends a notification to the team
What I would refine for a client
The weakest point is reading orders written in plain language in direct messages. A human is still needed to confirm those. In a deployment I would add a language model for a first pass at extracting line items, but with mandatory human confirmation, because a wrong quantity costs more than the minute it saves.
Do you have a similar process?
Describe it in a few sentences. I will tell you whether it can be automated and what it would cost — including when the answer is that it is not worth it.