Customer · Phase 2
Voice — Workshop
Capture tasks by voice, document hands-free. The mechanic has grease on their hands — and still everything's in the twin.
What it does
01
Push-to-talk mobile
One button, short voice input — task or doc entry lands in the twin.
02
Structured extraction
LLM extracts job, machine, step, time — no free-text rubbish.
03
Offline buffer
Recordings are stored locally and synced once you're back online.
How it runs
- 01
Record
Push-to-talk. Recording runs, buffer is set.
- 02
Transcribe
Speech recognition local or cloud, per tenant setting.
- 03
Structure
LLM produces a structured entry with field mapping. You only confirm.
What you get
- Structured doc entries without typing
- Tasks with the right context
- Voice log as audio for compliance
Requires
- → Tasks module, Workshop module.
In the example
–40 %
capture time for routine tasks
90 d
voice retention (default)