We did the *Start with Why* exercise at Darwin and it stuck. Our **what** is building an AI workforce that talks to a business's customers, on every channel they use, all the time. Our **how** is freeing humans from repetitive, boring work. Our **why** is *let humans do human things* — let humans spend their lives doing the most human work possible. 
   ╭─────────────────────────────────────────────╮
   │                                             │
   │   WHAT  ai workforce for mid-market         │
   │                                             │
   │   ╭─────────────────────────────────────╮   │
   │   │                                     │   │
   │   │   HOW   free humans from            │   │
   │   │         tedious tasks               │   │
   │   │                                     │   │
   │   │   ╭─────────────────────────────╮   │   │
   │   │   │                             │   │   │
   │   │   │   WHY   let humans do       │   │   │
   │   │   │         human things        │   │   │
   │   │   │                             │   │   │
   │   │   ╰─────────────────────────────╯   │   │
   │   │                                     │   │
   │   ╰─────────────────────────────────────╯   │
   │                                             │
   ╰─────────────────────────────────────────────╯
The reason this why matters more than it sounds is that it forces us to constantly ask: **what is human?** Today, looking someone in the eye and giving them your word — that's still human. Maybe in ten years there are hyper-realistic robots doing it convincingly; we don't think so, in the 5–10 year window. So that's a category we want to expand and protect. I read a book recently on animal evolution. Humans developed an unusual amount of mental hardware for facial recognition and movement reading. We are wired to read each other's faces in extraordinary detail, much more than we read anything else. That circuitry is going to be exercised more in the next decade than ever before, because it's the part of us AI can't replicate. We want to lean on it. We think it's *inhuman* to do something repetitive, identical, hour after hour, year after year. That's the work AI should take. The work that involves judgment, presence, care, a face — that's the work humans should keep. This frame has implications for what we build, who we hire, and how we evaluate ourselves. If a feature would automate a moment of real human connection — a hard conversation, a moment of trust, a goodbye — we hesitate. If it would automate the hundredth identical reply to *"what are your hours?"*, we ship it without flinching. The why isn't a poster in the office. It's a question we ask in every product decision: *is this making a human's life more human, or less?*