Workflow Automation
I build end-to-end automated processes that replace manual, repetitive work with reliable, self-running systems. Using tools like n8n, custom scripts, and orchestration platforms, I design workflows that trigger on events, coordinate across multiple systems, and handle errors gracefully.
Automation that actually works
The difference between a demo and a production workflow is error handling, monitoring, and the ability to evolve over time.
Most workflow automation tools make the first 80% easy and the last 20% painful. The initial setup is fast — connect a trigger, add a few steps, and it works in testing. But production is different. APIs return unexpected errors, data arrives in formats you didn't anticipate, and edge cases multiply. I build workflows with that last 20% in mind from the start.
I favour n8n for most workflow automation because it's self-hostable, open source, and gives you full control over your data and execution environment. But I'm tool-agnostic — if your team already uses another platform, or if a custom script is the right approach, I'll work with what makes sense for your situation.
What I build
Workflows that coordinate tools, handle complexity, and run without babysitting.
Trigger-based workflows
Workflows that fire automatically when something happens — a new email arrives, a form is submitted, a record changes in your database, or a webhook is received. The trigger starts a chain of automated actions with no manual intervention needed.
Multi-step orchestration
Complex workflows that coordinate across multiple systems and services. Data flows from one step to the next, with conditional branching, parallel execution, and aggregation of results from multiple sources.
Tool and service integration
Connecting the tools your team already uses — CRMs, email platforms, databases, cloud storage, messaging apps, and custom APIs. I build the glue that makes your tools work together as a unified system rather than isolated silos.
Error handling and retry logic
Production workflows need to handle failures without crashing. I build in retry logic with exponential backoff, fallback paths for when services are unavailable, dead-letter queues for failed items, and notifications when human attention is needed.
Monitoring and alerting
Every workflow I build includes observability. Execution logs, success/failure metrics, duration tracking, and alerting on anomalies. You should always know whether your automations are running, how they're performing, and when something needs attention.
AI-enhanced workflows
Workflows that incorporate LLM calls at specific steps — classification, extraction, summarisation, or decision-making. The AI handles the parts that need judgement while the workflow engine handles the deterministic routing and execution.
Low-code, maintainable by your team
I build on visual platforms like n8n wherever it makes sense, so your non-technical team members can inspect, tweak, and extend workflows without writing code. The goal is handoff — you own it and can keep it running without me.
Ready to stop doing things manually?
Tell me about the processes eating up your team's time. I'll design workflows that handle them automatically.