Less copying, fewer spreadsheets, nothing that fails quietly.
Copying data between tools, rebuilding the same weekly report, or relying on a script only one person understands? I can turn that into reliable Python software that works with your tools and tells someone when something goes wrong.
First step A short call, then a written proposal with a fixed price for a defined first piece.
What this looks like
- Orders, invoices or form submissions moved between systems through their APIs, with validation before anything is written.
- A recurring report built automatically and sent to the people who need it.
- A small internal tool with a simple interface, history and exports, replacing a shared spreadsheet.
- Documents or emails read and sorted with AI where the input is messy, with a person checking anything uncertain.
A good fit if
- Someone spends hours every week on steps that are always the same.
- Important data lives in spreadsheets or email threads.
- Two systems should talk to each other and currently do not.
- An existing script works, but nobody trusts it to run unattended.
Probably not a fit if
- You need a large public web application built from scratch by a team.
- You want a no-code workflow and your team already maintains one well. Keep it.
How I approach it
-
Start from the real process
I map how the work is done today, including the exceptions, before deciding what to automate.
-
Built to run unattended
Clear inputs, explicit failure handling, logs of every run, and alerts that reach a person.
-
AI only where it helps
Reading messy documents may need AI. Moving data between systems usually does not.
-
Yours to keep
Documented, tested code in your repositories, which your team can run and change.
Evidence
Questions
Is my problem too small?
Probably not. If a manual process costs your team real time every week, it is worth a conversation. Small, well-defined automations are often the best first project.
Do you use no-code tools like Zapier, Make or n8n?
If your team already uses one and it fits, yes. For logic that needs tests, versioning or custom integrations, plain code is usually cheaper to maintain.
Where does it run?
Wherever suits you: a small server, your existing cloud account, or a container on a platform you already have. I will suggest the simplest option that is reliable.