AI Productivity

What AI actually changed about how I run product.

AI is a multiplier, not a replacement. These are the workflows, prompts and team habits I use to cut repetition, sharpen thinking and ship more often with less stress.

Pick the model for the job

Some models are better at images, some at structured data, some at cheap high-volume work. A quick comparison of Claude, GPT, Qwen and Kimi, and a simple way to test them on your own tasks.

TensorX — sovereign inference

Sovereign AI is becoming one of the busiest use cases in product work. For most day-to-day tasks I keep the team on a single platform like Claude Code, but when data sovereignty matters or context continuity is less important, switching to TensorX gives you cheaper, private inference on EU infrastructure — and a clean way to compare outputs across models.

Jentic Arazzo UI — workflow visualisation

Multi-step AI workflows quickly become hard to read. Arazzo UI turns them into shareable diagrams so the team can see dependencies, handoffs and failure paths before they become surprises.