Home Luware Nimbus Usage of Nimbus Configuration Workflows Workflow Activities Conversation Handling Activities

Input Customer (Advanced)

Input Customer (Advanced) (IVR)

Description

Plays/shows a prompt while allowing the caller to perform voice and DTMF input.

☝ PLEASE NOTE

This action uses up to 10 exit nodes to react to customer input. When the corresponding prompt properties are not filled out, saving changes in this activity is not possible.

✅ PRECONDITIONS FOR VOICE INPUT

If you want to offer convenient voice input to your customers, you need to add and configure a Speech Recognizers (used for voice to text analysis and parsing).

💡This can be combined with other methods (e.g. Dialpad DTMF tones), e.g. when the user request is not recognized, or speech recognition services are temporarily not available.

Required Predecessor

License

Enterprise Routing

Contact Center

Modalities

Audio/Video

Instant Messaging

External Task

Properties

💡 Some of the fields/properties are only shown when the corresponding option is toggled on.

Configurable Properties	Description
Main Prompt Type	Text-to-Speech (with Language Select & Text Field) Audio Resources. 🔍 Audio resources can be added via Configuration > Resources.
Prompt Language	Allows to select the locale/voice to use when using the Text-to-Speech engine.
Prompt Text	Allows you to add text that will be read in the prompt language/locale. You can use custom and system parameters in text-to-speech prompts. This way, it is easier to guide your customers properly through IVR. To make use of the functionality, Add Parameters need to be enabled first. Once enabled, you can add custom and system parameters to your prompt text, e.g.:
Max Input Timeout hh:mm:ss	Maximum wait time for any caller response to be started. Default 00:00:30 Min: 00:00:03 Max 00:30:00 💡 Once started, the “Time to complete started voice input” / “Time to complete started DTMF input” field applies. 💡 If nothing was entered, the "No Recognition (NR)" exit node is used.
Max Number Retries Enabled (toggle)	When disabled: Only one recognition attempt is made. When enabled: A maximum number for retries can be defined. Default: 3 Min: 1 Max 99 Behavior If a maximum number “x” of retries is set, the NR (Not recognized) exit is taken if one of the following scenarios apply: The customer entered a wrong value “x” times The system went “x” times through the state "nothing received" (missing information¹). This also applies in case the customer asks repeatedly for an explanation. Any “Max Time” threshold is reached ¹💡Missing/wrong inputs are treated as described further down in the table under → “Missing Input”.
Voice Recognition (toggle)	Allows direct speech input for the customer when toggled on. ☝Note: Language detection accuracy, quality and speed depend on the configured Speech Recognizer. → Also see “Speech Recognizer" selection below.
Time to complete started voice input hh:mm:ss	Specifies the time the caller has to make the voice input. Default 00:00:02 Min: 00:00:01 Max 00:00:09
Speech Recognizer	The speech recognizer used for voice input processing. ✅ Precondition: Requires Speech Recognizers need to be configured via Configuration > Virtual Assistants > Speech Recognizers.
DTMF (toggle)	Allows DTMF input for the customer when toggled on.
Time to complete started DTMF input hh:mm:ss	Specifies the time the caller has to complete the next DTMF input. Default 00:00:02 Min: 00:00:01 Max 00:00:09
Input during announcement (barge in)	When enabled: … the caller is allowed to enter DTMF input the moment the main prompt was started. … the "Time to complete started input" timer starts immediately upon first input. Any prompts at this point are stopped. When disabled: … the caller is allowed to enter DTMF input only afer the main prompt has played in full. … the "Time to complete started input" timer starts after the main prompt. Voice Input: Limitation by Microsoft Barging in (using voice input during announcement) is currently not possible. This is a limitation from Microsoft and there is currently no workaround. → Regular “clean" DTMF input is not affected by this. We will update or remove this notice when suitable workarounds or improvements can be made.
Parameter to Save	Allows to store the customer's input (voice input or DTMF) to a custom parameter for further usage. To do so, you can chose the custom parameter you want to use for storing the customer's input in the dropdown Parameter To Save. When a parameter is selected, the Save on No Recognition toggle becomes available.
Save on No Recognition	Will store customer inputs in a custom Parameter even if the No Recognition (NR) exit is taken. If the Save on No Recognition toggle is enabled, the input is stored upon leaving the workflow activity through any of the exits. If the Save on No Recognition toggle is disabled, the input is only stored upon leaving the activity through any of the recognized exits (when matching the input with the expectation). General Limitation on voice inputs If there is a silence on the caller's end, any noise recorded might still be considered as a “wrong” or “missing” input. This is a design limitation that can be caused by the static noise or phone system shortcomings outside of Nimbus control. ☝ For your later parameter validation within workflows, keep in mind that checking this option will override the parameter default value with anything that was picked up. → If you regularly encounter issues with erroneous customer inputs, we recommend using the Nimbus Power Automate Connector to store customer inputs to be later checked by an Agent / Supervisor, e.g. by showing the parameter in My Sessions or storing it in a list for later workflow improvements.
Expected Input	Description
Name	A clear name for the exit node. ✅ You can add up to 10 nodes by clicking "Add Input" at the end this field group.
DTMF	Allows to enter multiple DTMF tones/digits per exit node. 🔍 If a user hasn't entered all expected digits yet – or the input is ambiguous (expected input 1: 123, input 2: 12, the caller enters "12") – the "Time to complete started input" applies before the DTMF is evaluated. 🔍 Unexpected key combinations are marked as as unclear and escalated with a “Wrong Input” prompt.
Voice Input	The voice input to be made by the customer. ☝Note: Only single words are matched in the “Expected Input” validation You can configure multiple expected keywords, by providing a comma-separated list of keywords per exit. There is no “fuzzy” logic to interpret word variants like “`Insurance, Insured`”, so you need to include them in a validation list as well. → Also refer to the best practice below for handling numerical and multi-language inputs. 💡Best Practice: Handling numerical voice inputs From the Microsoft Speech-To-Text documentation: Inverse Text Normalization (ITN) is a process that converts spoken words into their written form. For example, the spoken word "four" is converted to the written form "4". 💡The Microsoft Speech-To-Text engine interprets spoken words in a fixed manner, meaning that you might not get an expected input “1” if the customer says “One”, you best handle all possibilities as one expected input list, e.g. “`1, One`”. 💡When your Speech-Recognizer is set to be Multilingual, you need to cover those variants as a list, e.g. “`Eins, Uno, One`”. 🔎Virtual Users as AI-driven alternative: If you want to offer a more advanced interaction with the customer – including natural language conversation with intent detection – consider using Virtual Users in conjunction with the “Add Virtual User” activity.
Missing Input	Description
Timeout hh:mm:ss	Time for the caller to enter input. Afterwards the 1st prompt is played. Default 00:00:06 Min: 00:00:05 Max 00:00:15 💡 Note: For voice input, the maximum recording time is timeout -3 seconds. The system needs this time to process and analyze the input.
1st Prompt 2nd Prompt	Message to play when the input was not occurring in the timeout. A selection of the following options: Text-to-Speech (with Language Select & Text Field) Resources 🔍 Audio resources can be added via Configuration > Resources . If User enters no input within the silence timeout: 1st Prompt → (if configured) does the following: Play "1st Prompt" otherwise play the “Main Prompt Text” 2nd Prompt → (if configured) does the following: Play "2nd Prompt", Play "1st Prompt" otherwise play the "Main Prompt Text"
Wrong Input	Description
1st Prompt 2nd Prompt	Message to play when the input was not recognized. A selection of the following options: Text-to-Speech (with Language Select & Text Field) Resources 🔍 Audio resources can be added via Configuration > Resources . If User enters a wrong input: 1st Prompt → (if configured) does the following: Play "1st Prompt" otherwise play the Main “Prompt Text” 2nd Prompt → (if configured) does the following: Play "2nd Prompt", Play "1st Prompt" otherwise play the Main "Prompt Text"

Input Limitation

Voice input via PSTN call - Silence detection is currently not working for voice input when the calling from a PSTN. In this case, the maximum recording time is recorded. The maximum recording time is calculated as follows: missing input timeout -3 seconds (this time is needed to process and analyze the recording). Note that this is a limitation by Microsoft and there is currently no more optimal workaround possible.
DTMF input - It is not possible for the caller to use the hash symbol (#) when entering DTMF input.

Manual Replacement of Input Customer with Input Customer (Advanced)

By right-clicking on the Input Customer workflow activity, you can replace it with Input Customer (Advanced):

💡 This only works in case the workflow doesn't have a workflow template with workflow or structure lock.

When replacing Input Customer with Input Customer (Advanced), all connections (entries and exits) and all values from all properties are preserved.

💡 It is currently not possible to downgrade from Input Customer (Advanced) to Input Customer.

Luware Knowledge

Luware Nimbus

Luware Recording

Resource Library

Input Customer (Advanced)