Workforce Engagement Management

 View Only

Sign Up

Back to discussions
Expand all | Collapse all

Voice CSAT Survey - No way to disable speech recognition (ASR) and force DTMF-only responses

  • 1.  Voice CSAT Survey - No way to disable speech recognition (ASR) and force DTMF-only responses

    Posted 09-09-2026 23:18

    Hi all,

    We're running into a limitation with Voice Surveys in Quality Management and wanted to see if others in the WEM/Quality community have dealt with this.

    The issue: Our voice CSAT survey is registering false ratings when there's no real, intentional response from the customer. We've seen this happen in a few scenarios:

    • The call is answered by a voicemail/answering machine, and the greeting audio gets picked up as a valid rating
    • The call is answered by an automated assistant instead of a real person
    • The customer leaves the line open without responding, and ambient audio still gets interpreted as a rating

    In all these cases, the survey shows as "Complete" with a score, even though no customer actually pressed a key or intentionally answered.

    What we'd like: the ability to force the survey to accept DTMF (keypad) input only, disabling speech recognition entirely, so that only a deliberate button press counts as a response.

    There's currently no native setting to do this in Voice Survey flows. There's an existing enhancement request for it (still in "Future Considerations"): https://genesyscloud.aha.io/ideas/ideas/CESVY-I-50

    Has anyone found a practical way to enforce DTMF-only on a voice survey, or dealt with false CSAT ratings from non-customer audio (voicemail, IVR crosstalk, silence)? How did you handle it - configuration change, process workaround, or did you just exclude those interactions from reporting?

    Any input is appreciated!


    #QualityEvaluations

    ------------------------------
    Roberto
    Partner Integrator
    ------------------------------


  • 2.  RE: Voice CSAT Survey - No way to disable speech recognition (ASR) and force DTMF-only responses

    Posted 09-10-2026 11:00

    hi Roberto, 

    We've seen similar behavior in a few deployments. Unfortunately, as you mentioned, Voice Surveys currently do not provide a native option to disable ASR and enforce DTMF-only responses, so ambient audio, voicemail greetings, or automated systems can occasionally be interpreted as valid answers.

    A few workarounds we've used or discussed with customers are:

    • Review survey invitation criteria to minimize sending surveys to calls that may have been transferred to voicemail systems or automated endpoints.
    • Post-survey reporting filters to identify and exclude interactions with suspicious patterns (very short survey duration, known voicemail destinations, abandoned interactions, etc.).
    • Conversation-level analytics to detect responses that appear to originate from answering machines or automated attendants and remove them from CSAT calculations during reporting.
    • Custom survey approaches using Architect, where more control over input handling may be possible, although this adds complexity and still may not fully eliminate speech recognition behavior depending on the design.

    From a product perspective, I agree that a DTMF-only option would be valuable, especially for regulated environments or organizations that require an explicit customer action before accepting a rating. Until the enhancement request progresses, most customers seem to rely on reporting controls and exclusion rules for these edge cases.

    I'd encourage anyone experiencing the same issue to upvote CESVY-I-50, as the ability to require a keypad press would help improve CSAT data integrity and reduce false-positive survey completions.

    Curious to hear whether anyone has found a more elegant workaround in production.



    ------------------------------
    Cesar
    INDRA COLOMBIA
    ------------------------------



Related Content