Voice Commands
Voice-to-Workflow lets you describe the next automation step out loud and have the AI Copilot translate your words into a fully-configured workflow action — no typing required.
How It Works
- Open a workflow in the Designer
- Click the 🎙️ Voice button in the Recorder toolbar (or press
Ctrl + Shift + M) - Speak your instruction clearly — e.g., "Click the Login button"
- The AI analyses your words, detects the target element on screen, and suggests a workflow step
- Review the suggestion in the Suggestion Tray and click Accept, Edit, or Dismiss
Voice Command Examples
| You say… | Activity generated |
|---|---|
| "Click the Submit button" | Click with the Submit button selector auto-filled |
| "Type my email address into the username field" | TypeInto with the username field selector |
| "Get the text from the total price label" | GetText targeting the price element; output stored in a new variable |
| "Open Chrome and go to google dot com" | OpenBrowser(Chrome) + Navigate("https://google.com") |
| "Wait for the spinner to disappear" | WaitForElementNotVisible with spinner selector |
| "If the error banner is visible, log the message" | If → GetText → Log |
| "Loop through each row in the data table" | ForEachRow on the most recent DataTable variable |
| "Take a screenshot and save it to the desktop" | CaptureScreenshot with C:\Users\…\Desktop\ path |
Supported Languages
Voice recognition is powered by Azure Speech Services and supports:
- English (en-US, en-GB, en-AU, en-IN)
- Spanish (es-ES, es-MX)
- French (fr-FR)
- German (de-DE)
- Japanese (ja-JP)
- Portuguese (pt-BR)
Change the recognition language under Settings → AI → Speech Language.
Requirements
| Requirement | Detail |
|---|---|
| Microphone | Any USB or built-in microphone; headset recommended for noisy environments |
| Azure Speech Key | Enter your key under Settings → AI → Azure Speech Key |
| AI Backend / Azure OpenAI | One must be configured for intent analysis (see Copilot Overview) |
| Windows Permission | Allow microphone access when prompted by Windows |
Suggestion Tray
After you speak, the Suggestion Tray slides up at the bottom of the designer showing:
- Activity type — what action was detected
- Confidence score — how certain the AI is (shown as a percentage badge)
- Pre-filled properties — selector, input text, variable names already set
- Accept — inserts the activity at the current cursor position
- Edit — opens the activity in the Properties Panel for review before inserting
- Dismiss — discards the suggestion without inserting anything
Get Better ResultsSpeak in complete sentences including the target element name exactly as it appears on screen. For example: "Click the blue Save Changes button" works better than "Click Save" when there are multiple buttons on screen.
Privacy NoteVoice audio is streamed to Azure Speech Services for transcription. No audio is stored by Genzbots. If you work with sensitive information, use the text input in the Copilot panel instead.