Most companies have repetitive processes that consume hours of human work every week: copying data between systems, generating reports, sending notifications, synchronising platforms. All of that can be automated.
My approach combines low-code tools such as n8n and Make for straightforward workflows with custom development in Python or Node.js when the business logic is more demanding. Automation does not need to be expensive or difficult to maintain.
Typical stack
- n8n — Open-source workflow orchestration with full control.
- Make (Integromat) — Fast automation between SaaS tools.
- Python / Node.js — Custom scripts and services when low-code is not enough.
- REST APIs / Webhooks — Integration between almost any system.
Real projects
- Invoice-processing pipeline: OCR and GPT extraction, automatic validation and ERP registration. Four hours of daily manual work reduced to zero.
- Automatic customer onboarding: from the web form to account creation, credential delivery and environment setup.
- Two-way synchronisation between CRM, email marketing and support systems, eliminating duplicated data.
FAQ
- Which processes can be automated with AI and APIs?
- Documentation, reporting, CRM work, data classification, onboarding, tool synchronisation, invoicing and document review often have strong automation potential.
- Do you use low-code tools or custom development?
- I use both. n8n and Make are useful for fast workflows; Python, Node.js and custom APIs are better when the logic is complex or reliability matters.
- How do you avoid fragile automations?
- By designing logs, retries, validations, alerts and human review points when an automated decision has meaningful business impact.
If your team spends time on work a script could do, you are losing money. Contact me at [email protected].