AI-in-the-loop automation — n8n, Make and your own APIs

Most AI automation is a demo — one flashy workflow that never touches production. I run the other kind: a fleet of scheduled automations that move real data, generate real reports, and post real content, every day, for a live multi-brand operation.

n8n and Make workflows wired to real APIs — not toy integrations. Data moves between apps on a schedule. Reports generate themselves and land in an inbox. AI gets wired in exactly where it earns its place: summarizing, deciding, drafting — never bolted on because "AI" sells.

Behind it sits a Postgres/Supabase core every workflow reads from and writes back to, so nothing is a black box — every automated action is logged: what ran, when, and what it touched.

The same shape, scoped to your stack: n8n or Make workflows around your real tools, a database that's the source of truth instead of a spreadsheet quietly rotting, and scheduled jobs that remove the repetitive work without removing your visibility into what's happening.

This isn't just a "connect two apps" job — though I do those too, when that's genuinely what's needed. The systems I build are for when the busywork has enough steps and enough stakes that you want it logged, scheduled and reliable, not automated once and forgotten.