Before starting, please make sure:

  • You've successfully added an Expert or Custom plan with CloudTalk

  • CloudTalk Support has enabled Speech-to-Text for your account

Setting up Automation

You will start by logging into your CloudTalk Dashboard. On the left side of the screen you will see a list of other tabs under Dashboard. To reach the automations tab, click on Account > Workflow Automations.

Click the blue button to + Add Workflow.

You can give your new Workflow Automation a name to distinguish it from other automated tasks you set up.

Exporting Speech-to-Text Transcriptions

In this example, we will actually need to use two Workflows: one to automate transcription, and one to automate exporting those transcriptions when they are finished. These will combine to form just one of many ways to automate workflow with CloudTalk.

Example Goal:

  • Forward all transcriptions to a separate CRM or management system account.


  1. Start by defining a trigger. In this example, our trigger will be whenever a recording gets uploaded. Under Object, we will select recording. For Action, we select uploaded.

    Note: This is the only available action for a recording Object.

  2. We could choose to set up an optional condition. For the purposes of our example, though, we don't really need one. Instead, click the triple dots in the righthand corner of the Conditions box to delete.

  3. The next step is to choose an action which will happen whenever a recording gets uploaded. We select the Action type Speech to Text.

    1. For Key>language, enter into the Value box the appropriate language for our recordings. Choose from the Google Cloud list of languages and copy/paste the code listed in its BCP-47 column.

    2. For Key>model, you can choose a specific machine learning model. In most cases the phone_call model is sufficient for the CloudTalk platform.

    3. When Key>enhancedModel is set to true, a more advanced model which has been optimized for accuracy will be applied. However, this option will increase costs of transcription.

  4. Confirm and continue, Save and Activate.

  5. Now we must make a second Workflow for automating the export. It is always advisable to name Workflows descriptively.

  6. Define a trigger of Object type transcription and Action type created. This trigger will initiate any time a transcription gets created.

    Note: This is the only available action for a transcription Object.

  7. There are some relevant tags to choose from if Conditions were necessary, though we do not need any in our example. Click the triple dots to delete the Conditions field.

  8. As an Action type, we choose to make an API request. For this kind of action, some information will be needed from the CRM system we want to send this information to. The first piece of info necessary is the Endpoint, which will look something like the fake example in the picture below. It will be some sort of API or token URL. As a Method, we choose POST for our example, as we will be sending or posting something to the desired endpoint.

  9. There are two subsections: one named Headers and one named Values, and these can take multiple [Key, Value] pairs. In our example, we want to keep things a bit simple. Under the Values subsection, we make the Key>transcription while we make the Value>{{ }}. The latter is a dynamic value which populates the field with the transcription in question, converting it to a CSV format to send to our receiving management system API.

    Note:  Additional entries in the Headers or Values subsections
    will vary depending on the requirements expected on the side of
    the receiving API. For example, there are usually API Keys or
    other tokens expected for authorization in the Headers subsection, as well as a Header for content type.

Once we save and set our Workflows to active, we will have accomplished our goal of forwarding all transcriptions to another system automatically, a process which we can later edit and toggle off or on at our choosing.

Troubleshooting Tips

  • For information regarding transcription, refer to Google's list of languages, transcription models, and advanced transcription models.

  • In most cases, sending information to another API (such as one associated with a CRM or other internet-connected tool) requires one or more keys or authorization tokens. Ask an admin associated with the receiving API if you have trouble finding this information.

  • Faulty conditions are often the result of unrecognized inputs in the Property box, or inputs in the Value box which will always result in no matches.

  • After clicking into the Property box, you must click one of the pre-selected keywords within the Useful data for your Workflow section to the right.

  • You can also refer to the Useful data for your Workflow section to validate whether the Value you entered seems to be a good match for the property you've chosen.

Related Links

What is a Trigger?

What are Conditions?

Did this answer your question?