AI marketing: what to automate today (and what not)
Not everything should be automated. But what should be, must be done well.
It's not hype, it's leverage
AI in marketing doesn't replace people. It frees them from repetitive tasks that burn hours without creating value.
The 3 practical levels
Level 1 — Generation
Content drafts, first brief versions, first ad copy versions. Always human-reviewed.
Level 2 — Workflow
Automatic lead qualification, scoring, routing. Agents that pre-process and bring only warm leads to your calendar.
Level 3 — System
Permanent agents that monitor competitors, market, trends, and produce weekly insights.
What NOT to automate
- Conversations that close big budgets
- Creative direction on hero projects
- Strategic decisions
- Crisis reactions
How to start
- Map your marketing processes by hours/week
- Identify the 3 most repetitive, low-value ones
- Prototype an agent for the first
- Measure recovered time
- Iterate
The AI competitive advantage isn't 'using ChatGPT'. It's having processes.
25 years dismantling businesses, finding the bottleneck and putting them back on track. E-commerce, SaaS, B2B, local: he's seen it all and has the right scar for every industry. He lands his clients between €10k and €50k of new revenue in the first 6 months — not by 'believing in awareness', but by designing funnels that work and measuring them every day.