AI Evangelist
Cobus Greyling
I write and speak on the intersection of language and AI. AI Agents, Harness Engineering and enterprise adoption.
In the press
Quoted on how conversational systems behave, and on what enterprises need before they trust an agent.
- AITech3652026The model is rented. The harness is owned
- MIT Technology Review2026Moltbook was peak AI theater
- Fast CompanyQuotedTrust, bounded autonomy, and sandboxed execution
- R&D World2026The real race is distribution, not models
- R&D World2026OpenAI, OpenClaw, and Moltbook
- MarTech360InterviewInterview with Cobus Greyling of Kore.ai
- Forbes2022Turn-taking, and why conversational AI is the interface
Latest writing
Five recent pieces from Medium. The same essays go out on Substack.
- Medium29 Oct 2026Governing enterprise AI traffic at scale
- Medium28 Oct 2026NVIDIA Nemotron 3.5 Transcribe
- Medium27 Oct 2026Did Qwen3.8 Really Try to Phone China? What Cohere’s Sandbox Traces Actually Show
- Medium26 Oct 2026Harness-Aware Distillation Demystified
- Medium23 Oct 2026There is a disconnect between what agents do and what they say they do
Talks and bylines
- Show reelVideoCobus Greyling, on stage
- VUX World2026Why intents, graphs and conversation design have come full circle
- TechRadar2026Enterprise AI governance cannot live in a prompt
- TechRadar2026How to build AI agents that don’t break at scale
- UX Magazine2022Age of Invisible Machines
- UX MagazineFellowContributor page, three pieces
- VUX World2022Talking conversational AI