AI Speech to Text
The AI Speech to Text feature allows the bot to automatically recognize and extract content from user speech (via an audio file), serving the purposes of storage, analysis, or automating internal data processing. Unlike typical question-answer features, the result after conversion is not displayed back to the user, but is forwarded to an analysis system, such as Mautic or CRM.
Real-life Applications:
- Analyzing User Feedback from Speech:
Users send audio files with feedback, comments, or questions.
AI converts to text → sent to the CRM system for sentiment analysis, content categorization, and priority assessment.
- Call Content Summarization for Customer Care:
After a call, the recording is sent to the system.
AI converts it to text for storage & call quality assessment.
- Automating Internal Processes:
Example: Employees record quick reports when it's inconvenient to type.
The bot converts speech to text → sends it to the internal processing system.
Example:
User:
Send audio file with feedback
Bot (AI Convert Speech to Text):
Content: "I am very satisfied with the service today, the staff was enthusiastic and provided great support."
Setup Steps:
Step 1: Create Basic Blocks
- Block “question”
Receive the audio file from the user and save it to the {{question}}
attribute.
- Block ChatGPT Convert Speech to Text
Create a new block → Integrate AI → ChatGPT
Type: Convert speech to text
Message send to OpenAI:
{{question}}
- Save response value to attribute:
{{answer}}
- If processing is successful:
→ Text {{answer}}
is sent to the Mautic system or stored in CRM for storage/analysis.
- Block “failed”
If processing fails → The bot redirects to the failed
block to display an error message.
Flow Diagram Overview:
- Flow:
User records or speaks → saves to
question
Sends
question
to ChatGPT → recognizes speech and converts it to text → saves toanswer
If successful → moves to the
Mautic
block for further processing (send email, store data, etc.)If failed → redirects to
failed
block
Tips:
You can combine with additional content analysis steps (e.g.,
sentiment analysis
,keyword extraction
, etc.) after receiving the text from AI.Ideal for internal bots, consulting chatbots, or post-sales customer service solutions.