DTMF Response - Voice Node Summary

Purpose and Use Cases

• The DTMF Response - Voice node is designed to collect input from customers during voice interactions, using Dual-Tone Multi-Frequency (DTMF) signals generated by keypad presses or speech input.
• It is used for menu-based navigation (e.g., “Press 1 for order status, Press 2 for support”), collecting fixed-length data (like a 10-digit mobile number), variable-length data (such as order IDs followed by an end key), and sensitive information (PCI DSS and PII data like credit card numbers, CVV, PIN, OTP, social security numbers, and date of birth).
• The node can be configured to accept DTMF, Speech, or both, with speech input options available for enhanced flexibility.

Configuration Steps

• Enable the Voice toggle and select the required language in the Dialogue Tree settings.
• Add the DTMF Response - Voice node at the desired point in the dialogue.
• Configure the response type (DTMF, Speech, or both), number of input digits, valid digits, process variable (for storing input), and speech modal (for speech input).
• Set up timeout values for DTMF and speech, add end key for variable-length input, and define the number of retries for invalid or no input.
• Enable sensitive data collection options for PCI or PII compliance, ensuring proper masking and secure handling.
• Add custom messages for timeout and invalid input scenarios, and configure a default path to handle cases where the retry limit is exceeded.

Best Practices

• Keep prompts clear and concise, limit menu options to reduce confusion, and clearly state valid input options.
• Use fixed-length input only when the expected input length is known.
• For sensitive data, ensure the environment is properly set up for PCI compliance and data masking.

Recent Enhancements

• The DTMF node now uses the same voice as the configured Voicebot, ensuring a consistent voice experience without switching to the provider’s default voice.
• Audio variable support allows playback of audio files from URLs stored in variables, simplifying flow design and reducing latency.
• Detailed latency metrics for ASR, Bot, and TTS stages are now available in reporting dashboards for improved monitoring and troubleshooting.

Challenges

• Speech input may require specialized models and hints for accurate transcription, especially for numeric or command-based responses.
• Sensitive data collection requires proper environment setup and compliance with security standards.