Equipment Hall·Input Mechanics·10 min

Verbal Velocity

Stop typing. Start thinking out loud.

Goal

You use your voice as the primary interface to Copilot — moving faster by talking, not by crafting.

Why this matters

Typing forces you to edit while you think. Talking lets you think first and refine later. Speed beats syntax — clarity emerges while you speak, not before.

The drill

Use Copilot’s voice capability, where available in your environment, and shift the thinking work into talking:

  1. Talk before you know the answer. Just start speaking and let Copilot find structure while you talk.
  2. Stream, don’t script. Share context as it comes to mind. Add information out loud — instead of translating it into perfect sentences.
  3. Let Copilot do the polishing. Speak messy and incomplete, then have raw thinking turned into a structured result.

Make it this week’s habit: whenever you catch yourself typing more than two sentences, stop and ask: “Could I say this faster than I can type it?” Then switch to voice.

Extra round

One full Copilot session this week voice-only: no keyboard, no editing, no perfect phrasing. Just talk until the thought is out.

Micro-win

Today you spoke a thought before it was finished — and it became clear while you talked.

Remember

Speaking doesn’t replace thinking — it frees it from simultaneous editing. You still check the result before anyone sees it, as always.

Back to the hall →

FAQ

Why should I talk to Copilot instead of typing?

Because typing forces you to edit while you think. Speaking separates the two: think first, refine later — which makes you faster and often makes the thinking clearer.

Do I have to phrase things perfectly when speaking to Copilot?

No — that's the point of the exercise. Speak messy and incomplete; Copilot structures the raw thinking into a usable result afterwards.

When does voice input pay off most?

For anything longer than two typed sentences: braindumps, context descriptions, first drafts. The rule of thumb: if you can say it faster than type it, say it.

What if voice isn't available in my setup?

Availability depends on environment and licence — check what's active in your tenant. You can train the principle typed as well: dump it in raw, have it structured, instead of polishing upfront.

Won't the result get worse if I speak unstructured?

The raw input gets messier, the end result usually better — because more context arrives. The key is the second step: explicitly having Copilot structure and condense.