Lisa Esterhuizen

I build AI automations that ship into daily work—used live by executives on enterprise accounts, not stuck in slide decks—and I design them for handover, so non-technical teams can run what I ship. I’m remote and async by default, and I love turning messy inputs into reliable output teams can act on.

As an AI Implementation Specialist at RAY AI, I embed as the AI implementation and executive layer for the CEO of Senken (Berlin carbon-markets SaaS). I won RAY AI’s internal AI & automation competition with the CWN subtitle pipeline—one-pass AssemblyAI transcription from a watch folder, then a glossary-driven Claude cleanup that fixes misheard terminology while preserving timestamps—unlocking a one-window publish flow.

I also lead practical, production-minded enablement: champion finder (deployment & extension) to map sustainability decision-makers into HubSpot, an EA toolkit with meeting briefs and strict anti-AI-voice post drafting, and travel coordination tools that reduce long itineraries into a simple selection-to-itinerary flow. On the content side, I’ve owned CWN production (multi-platform, bilingual EN/DE subtitles) and built “Podcast Painkiller,” an AI-powered agent that edits transcripts, cuts video against the revised transcript, generates timestamped subtitles, and auto-uploads the finished episode.

Alongside implementation, I consult on end-to-end automation builds (e.g., Make.com workflows using Claude web-search agents and skills, plus structured JSON pipelines and approval gates). Earlier, I built a heart-first meal planner “with Claude” and worked as a hospital diet technician in New Zealand, so I’m comfortable with accuracy pressure, careful stakeholder handover, and clear technical writing.