What is VoiceBase?
Analyzing your phone call recordings can provide valuable insights for your campaign, however it can be extremely time consuming if done manually. VoiceBase is a web-service that analyzes audio, transcripts the dialogue and presents speech analytic results. Using VoiceBase, you can automatically transcribe your calls, scan for client sentiment and find frequently used keywords in your phone calls.
Integrating VoiceBase using Zapier
To automatically analyze call recordings, we will require 2 separate webhook tasks.
- The first webhook will be used to send call recording URLs to Zapier which forwards it into VoiceBase for parsing.
- The second webhook will be fired once VoiceBase is finished parsing the call recording and can store the results in a text file at location of your choice.
Zapier is a web-service that acts as a middle man to pass data from thousands of web-applications. Read more about the benefits of using Zapier here.
We will be integrating Retreaver and VoiceBase in this article however there is also an official VoiceBase Zapier Guide here: https://www.voicebase.com/zapier-integration/
Before we begin: Make sure you create a VoiceBase account to request a unique API key and also create a Zapier account to enable third party integrations. VoiceBase Account confirmations could take up to 48 hours.
Note: You must request an invite from VoiceBase from its main site for its application to be visible inside of Zapier as an action step, click here: VoiceBase Invite.
Firing off a webhook when a call is finished recording
- Log into Zapier and create a new zap, choose the webhook trigger option, this will provide you with a Zapier link that will listen for any incoming webhooks.
- Under your Campaign edit menu inside of Retreaver, create a new webhook that always fires after all other pixels fired and paste in the Zapier provided listener, then append the "?CallRecordingURL=[call_recording_url]" keyword at the end of it. Update your campaign. Note* [ Ensure Call Recording is Enabled under Campaign Toggles ]
- Create a test call recording by calling your campaign and speaking for a few seconds or alternatively, find a per-existing call recording in your call log by selecting a call UUID of your choice and scroll down to "Voice Recording" under the "Call Overview" menu. Lastly, manually enter your Zapier webhook and the audio url into your browser as so:
https://hooks.zapier.com/hooks/catch/XXXX/YYYY/?CallRecordingURL=https://callpixels.s3.amazonaws.com/recordings/XXXXX-XXXXX/embed/call_XXXXX.mp3
Once a sample call has been sent, press "ok i did this" on the Zapier page to continue. - Now we add the Voice Base Speech Transcription and Analytics Application as an Action that will be performed whenever Zapier triggers a new incoming audio file.
- Follow the prompts to authenticate your VoiceBase account with Zapier. To do so you will need to create a unique bearer token. Sign into the VoiceBase Platform and navigate to the bearer token management section. Create a new bearer token and click "copy to clipboard" before pressing "done"
- Back into Zapier, paste your copied token when it prompts you to do so to authenticate your account:
- Once authenticated, you will need to configure VoiceBase to receive the audio URL you forwarded to Zapier earlier under the "Media" and "Filename" fields.
You will also need to place the following code snippet under the VoiceBase configuration field. The URL is a temporary webhook placeholder that instructs VoiceBase where to report back once its finished processing, which you will receive and fill out in your second zap in later steps.
{ "speechModel" : { "language" : "en-US" }, "publish": { "callbacks": [ { "url" : "https://hooks.zapier.com/hooks/catch/PLACEHOLDER/REPLACE-ME/", "method" : "POST", "include" : [ "transcript", "knowledge", "metadata", "prediction", "streams", "spotting" ] } ] } }
There are also additional configuration options that can be used to tweak your VoiceBase behavior and final transcription output. For now we will be sticking to the defaults but VoiceBase does offer a helper tool for advanced users. - You will then be prompted to run a test of your new Zap, depending on a variety of third-party factors you might experience some trouble shooting errors. If that is the case, feel free to press "Skip Test" as the Zap will still work even though you skipped the test. Be sure to name and activate your newly created Zap.
Firing a webhook once VoiceBase Parses the Call Audio
In the previous section, we created a zap to fire whenever a call recording url is forwarded to Zapier which then sends the recording to VoiceBase, however we also need to create an additional Zap to fire once VoiceBase has finished processing the recording. This zap will fire a webhook back to Zapier to deliver the parsed information once its been processed.
- Begin by using this link invitation to create your new Zap using VoiceBase as the trigger.
- You will be prompted to sign into Zapier if you haven't already, and you will be notified that this Zap will be fired once VoiceBase finishes processing.
- You will be prompted to choose your VoiceBase account again, if asked to create an additional bearer token, follow step 5 & 6 from above and paste the token back into Zapier.
- You will be given a callback URL that must be placed in the configuration portion mentioned in step 7 above. It is under your first Zap-> Action Step-> Edit Template. This is the webhook that will be fired from your first zap to instruct VoiceBase to run Zap2 once its done processing.
- Once you press "ok i did this", Zapier will attempt to listen for a webhook to be sent to the URL you were just provided. You should attempt to call your campaign now and fire off a job into VoiceBase using your first Zap to confirm it is working, otherwise select a default sample or skip the test.
- Now that VoiceBase is successfully sending information to Zapier, its time to create an Action. In this example we will be storing the parsed results into a Google drive using a text file. Select the Google drive account you wish to use and press continue.
- For this example we specify a folder where the transcript will be created and stored in Google Drive. We named the transcript by the original recordings filename and we chose the text version of the transcript from VoiceBase as the files content [ There are multiple fields we can include, to be covered in an advanced guide later on ]. Finally we select "yes" to convert the transcript into an editable document.
- Skip the sample test or use your provided sample to ensure everything is working smoothly, name your zap and be sure to activate it.
In the first section, we created a Zap webhook to fire when a call recording was received which would then send the information to VoiceBase for parsing.
In the second section, we created a Zap webhook that would listen for a callback from VoiceBase indicating it finished parsing audio, which then triggers an action to store the results in a text document on a Google drive folder of your choosing.
Now each call recording is automatically sent to VoiceBase, evaluated and stored into a Google Drive Document. Note that Zapier is very customizable and you have the option of sending or moving data to a variety of web applications. You may instead choose to send results to a spreadsheet, CRM or any other web-service of your choice.
If you are having trouble integrating VoiceBase with Retreaver, feel free to contact support anytime.
Comments
Please sign in to leave a comment.