Work with voice survey flows

Prerequisites

One of the following licenses:

  • Genesys Cloud CX 1 WEM Add-on II, Genesys Cloud CX 2 WEM Add-on I, Genesys Cloud CX 3

Edge and Media Tier version:

To access published survey forms when you create a voice survey flow:

  • Quality > Survey Form > View permission

Notes:
  • You cannot use Genesys TTS as the TTS engine in a voice survey flow. You can use Genesys Enhanced TTS or a compatible third-party engine. For more information, see About text-to-speech (TTS) engines.
  • You can use the Genesys Enhanced TTS engine free of charge in voice survey flows. Genesys does not bill you for using the Genesys Enhanced TTS engine if you only use it in voice survey flows. To use the Enhanced TTS engine, you must install it first. For more information, see Install Genesys Enhanced TTS from the AppFoundry.
  • When you create a voice survey flow, Architect tries to select a compatible engine for the default language of your flow based on your organization’s settings. In other words, if the organization’s default TTS engine for the selected language is incompatible with the voice survey flow type, Architect tries to select a compatible TTS engine that is available for your organization. For example, if your administrator specified Genesys TTS as the default TTS engine for en-us, but you also have the Genesys Enhanced TTS engine installed in your organization, Architect selects the Genesys Enhanced TTS engine in the voice survey flow.
  • The maximum length of a flow name is 200 characters and it cannot contain a slash / or end in -debug.
  • You cannot duplicate names for voice survey flows.

Create a voice survey flow

To create a voice survey flow, follow these steps:

  1. From the Architect home page, click or hover over the Flows menu and select Voice Survey.
  2. Click  Add. The Create Voice Survey Flow dialog box opens.
  3. In the Name field, enter a unique name for the flow.
  4. (Optional) In the Description field, add a short sentence or phrase to describe this flow.
  5. Click the Survey Form list and select the survey form based on which you want Architect to generate the voice survey flow.
    Notes:
    • Architect currently displays the survey forms that have been published under Admin > Quality > Survey Forms in batches of 100.
    • You can only create a voice survey flow based on the latest published version of a survey form that has not been disabled.
    • The following limitations apply to survey forms:
      • As Architect creates a list slot type for each Multiple Choice, Yes/No, and Range question, the survey form can only contain a cumulative total of 50 questions of these question types. For example, if the survey form includes 20 multiple choice questions and 10 yes/no questions, Architect allows only 20 range questions in addition. There is no limit for questions of the Free text and Read only types. For more information about the question types, see Add questions to a survey form. For more information about the list slot type, see Slots and slot types overview.
      • The survey form can only contain one question of the Net Promoter Scoresm type.
      • Answers in the survey form can be up to 50 characters in length.
      • Questions of the Multiple Choice, Yes/No, and Range types can have a maximum of 20 answer options. Best practice recommends that you limit the number of answer options in a voice survey to fewer than 20.
      • If the voice survey only has Read Only questions, Architect ignores DTMF inputs and voice barge-in for the entire survey flow. If a question of another type follows a Read Only question, and the customer provides voice or DTMF input, then Architect skips the Read Only question and uses the input for the question that follows.
      • Architect currently limits the number of characters that transcripts of customer voice inputs can contain. If a transcript of a voice input exceeds 500 characters, Architect truncates the input it received as an answer. This limitation is most relevant for Free text questions and for the follow-up questions of NPS questions.
      • Long periods of speech recognition may cause issues for certain speech-to-text (STT) providers.
      • For questions that have an answer option of N/A (not applicable) in the survey form, the voice survey participant can only give a DTMF input as an answer. Architect assigns to such answer options the DTMF character value of star (*) and the communication value of "Press star for not applicable" by default. You can edit both values in Architect.
      • If the selected survey form contains any character encoding (for example, Unicode characters) or markup language (markdown) that you can use in your web surveys, be sure to edit the survey form and remove them. Encoded characters and markup language can be challenging for speech recognition and text-to-speech systems as they might not accurately interpret it or translate it into spoken language. If your survey form contains encoded characters or markup language and you want to use the same form for both web and voice surveys, Genesys recommends that you create a duplicate of the original form without the markup for the voice survey.
  6. Click the Default Language list and select the flow's default supported language.
    • You can select any available default language in Architect for any survey form, regardless of the language that the form was created in. In other words, at survey flow creation, you can select a default flow language that is different from the language you selected in the Default Survey Language field in the survey form. For example, if the survey form's default language is French (Français), you can select French Canada (fr-CA), but also German (de-DE).
    • Architect shows a warning message at flow creation if there is a mismatch between the survey form's language and the language you select as the voice survey's flow default language.
      Note: Architect issues the warning message as a helpful notification that unexpected compatibility issues might occur during runtime regarding speech recognition and text to speech.
       
    • Architect considers the language pairing a mismatch based on the following rules:
      • If the base language (without locale) of the survey form and the base language of the voice survey flow do not match, then Architect throws a warning. For example, when the survey form's language is French ("fr"), and you select German Switzerland ("de-CH").
      • If Architect does not display the regional variety (locale) of the survey form's language, then Architect does not throw a warning for any corresponding language locale of the base language. For example, if Architect shows the survey form's language as German ("de") or French ("fr"), all regional varieties (locales) of German or French are considered matching.
      • If Architect displays the regional variety (locale) of the survey form's language, then Architect throws a warning if the survey form's language locale and the default language's locale do not match. For example, if Architect shows the survey form's language as Portuguese Brasil ("pt-br") and you select Portuguese Portugal ("pt-pt").
        Note: English is an exception. If you select an English language locale other than the survey form's locale, Architect does not throw any warning.
    • If you select a language for your flow that is different from the survey form language and is not a locale of the language, speech recognition and text to speech might not work as expected. Also, communication texts might not be prompted or synthesized correctly.
  7. Click the Divisions list and select the division in which to place the flow.
  8. Click Create Flow. The flow's configuration page opens.
    Note: You cannot import a valid voice survey flow from another organization because the survey form ID will not be compatible.

Design logic behind voice survey flows

Architect automatically generates your initial voice survey flow based on the survey form that you selected with the following design logic:

  • Architect converts the Header and Footer survey form sections into Communicate actions.
  • Architect adds a Call Task action for each question group. If the survey form author chose to include a question group conditionally on the survey form, the Call Task action contains two paths: “Not Applicable” and “Default.”
  • Architect adds a Disconnect action after the Call Task actions to conclude the flow.
  • Each question group in the survey form becomes a separate Reusable task that the Call Task actions call. The task has the name of the specific question group. For more information, see Add questions to a survey form.
  • Architect uses the Ask Survey Question action for the individual questions within question groups.
  • If the survey form author chose to include the question group conditionally, Architect adds a Decision action at the start of the Reusable task that contains the question group. The Decision action uses an expression that Architect derives from the conditions listed in the survey form under Logic. Architect arranges the Ask Survey Question actions in the “Yes” path of the Decision action and adds an End Task action for the “No” path.
  • Architect adds an End Task action at the end of every Reusable task. The action takes the default output path.
  • Architect generates the following resources for the voice survey flow, depending on the question types in the survey form:
    • A slot for every question that is not a read-only question. Architect uses the following naming convention for generated voice survey slots:
      • Group{x}_Question{y} for multiple choice, range, yes/no and free text survey questions.
      • Group{x}_Question{y}_Follow_Up for follow-up questions related to an NPS question.
    • A list slot type for every multiple choice question (including range and yes/no survey questions). Architect uses the following naming convention for generated list slot types: Group{x}_Question{y}_Type. Slots for free text and NPS questions receive the built-in:any slot type.
  • Architect displays a warning message in the following scenarios:
    • If the default language you select for the voice survey flow does not match the survey form’s language.
    • If the survey form has been disabled since the voice survey flow was created.
    • If a new version of the survey form has been published since the voice survey flow was created.

    Customize your voice survey flow

    You can customize the following parts of the generated voice survey flow:

    • Change the generated name of any action
    • Change the generated communication output values for these areas:
      • Communicate actions (including the Survey Header and Survey Footer)
      • Question and Answer texts in Ask Survey Question actions
      • No Input, No Input Apology, and No Input Confirmation Apology communications
      • No Match, No Match Apology, and No Match Confirmation Apology communications
      • Confirmation Rejection Apology communication
    • Change the generated DTMF settings for these areas:
    • Change the default Timeout settings for these areas:
      • Inter-Digit Timeout
      • Voice No Input Timeout
      • Voice End of Input Timeout
    • Add new actions
    • Remove or reorder actions, including Ask Survey Question actions
    Note: You can create new slots of the generated custom slot types, but you cannot change any generated slots.

    Built-in variables

    Name Type Collection? Read Only? Description
    Flow.IsTest Boolean No Yes Indicates whether or not the flow is running in debug mode.
    Flow.StartDateTimeUtc DateTime No Yes The UTC datetime when the flow started execution.
    Flow.Version String No Yes The version of the active flow.
    Form.FullSurvey SurveyForm No Yes The survey form associated with the voice survey flow.
    Form.QuestionGroup{x} SurveyQuestionGroup No Yes The specific {x} question group in the survey form associated with the voice survey flow. 
    Form.Group{x}_Question{y} SurveyQuestion No Yes The specific {y} question in the specific {x} question group in the survey form associated with the voice survey flow.
    Session.ChannelType String No Yes

    The channel type of the conversation associated with the voice survey flow instance. It has one valid value: 'call'.

    Session.ConversationId String No Yes

    The ID of the conversation associated with the voice survey flow instance.

    Session.InvocationData String No Yes

    The invocation data that the Set Post-Flow action provides for the voice survey flow it launches. If no invocation data is available, this returns a NOT_SET JSON value.

    Session.Language String No Yes

    The lowercase IETF language tag for the language in which the flow is currently running. If there is no current flow language, this returns a NOT_SET string.

    Session.LastCollectionConfidence Decimal No Yes

    The participant's most recent utterance confidence in an Ask Survey Question action, Ask for Slot action or Ask for Yes/No action, excluding those for built-in confirmation or disambiguation questions. If no such utterance exists, then this is a NOT_SET decimal.

    Note: This variable applies specifically to automatic speech recognition (ASR) confidence, not natural language understanding (NLU) confidence.

    Session.LastCollectionUtterance
    String No Yes

    The participant's most recent utterance in  Ask Survey Question action, Ask for Slot action or Ask for Yes/No action, excluding those for built-in confirmation or disambiguation questions. If no such utterance exists, then this is a NOT_SET string.

    Session.Source String No Yes The type of application or mechanism that the end user employs to communicate with the voice survey flow. It has one valid value: 'phone'.
    System.MaxDate Date No Yes Holds the maximum value allowed for a Date, which is December 31st, 2200 ("2200-12-31").
    System.MaxDateTime DateTime No Yes Holds the maximum value allowed for a Date Time, which is December 31, 2200 at 11:59:59 p.m.
    System.MaxDuration Duration No Yes Holds the maximum value allowed for a Duration, which is 999999999999999 milliseconds (11574074 days, 1 hour, 46 minutes, and 39.999 seconds).
    System.MaxInt
    Integer No Yes Holds the maximum value allowed for an Integer, which is 999999999999999.
    System.MaxTime
    Time No Yes Holds the maximum value allowed for a Time, which is 11:59:59.999PM ("23:59:59.999").
    System.MinDate
    Date No Yes Holds the minimum value allowed for a Date, which is January 1st, 1800 ("1800-01-01").
    System.MinDateTime DateTime No Yes Holds the minimum value allowed for a Date Time, which is January 1st, 1800 at 12:00:00AM ("1800-01-01T00:00:00.000Z").
    System.MinDuration
    Duration No Yes Holds the minimum value allowed for a Duration, which is -999999999999999 milliseconds (-11574074 days, -1 hour, -46 minutes and -39.999 seconds).
    System.MinInt
    Integer No Yes

    Holds the minimum value allowed for an Integer, which is -999999999999999.

    System.MinTime Time No Yes

    Holds the minimum value allowed for a Time, which is 12:00:00AM ("00:00:00.000").


    Net Promoter, NPS, and the NPS-related emoticons are registered trademarks, and Net Promoter Score and Net Promoter System are service marks of Bain & Company, Inc., Satmetrix Systems, Inc. and Fred Reichheld.