All points of view
03 of 07
Intelligence is infrastructure.
Treat AI like electricity or the cloud — plan it, budget it and measure what you get back.
In plain words
AI costs money every time it is used, and different AI models cost very different amounts. So AI should be managed like other infrastructure: pick the right model for each job, give each team a budget, and check the value you get back.
Why Jenks thinks this
- As AI Lead at Babylon Labs he owns the AI infrastructure and the AI operations budget for the whole company.
- He asks hard money questions out loud — for example, when does an AI budget of about $20k a year per developer actually pay back, and how should spend be split between teams and models?
What he does about it
- Sends each task to the cheapest model that is good enough ("model routing").
- Sets a budget per team, and tests quality before a wide rollout.
- Tracks the return: time saved, how fast work gets done, and quality.
What it means for you
Before you scale AI, decide who owns the AI budget and how you will measure its value. Without that, costs grow quietly and nobody can say whether it worked.