detailed amazon feedback with uploaded call recordings

Post call analytics for your contact middle with Amazon language AI services

This blog post was last reviewed and updated March, 2022. Added support for Transcribe Call Analytics call summarization. See New features.

Your contact center connects your business concern to your community, enabling customers to order products, callers to request support, clients to make appointments, and much more. Each conversation with a caller is an opportunity to learn more about that caller'due south needs, and how well those needs were addressed during the call. Y'all tin uncover insights from these conversations that assistance y'all manage script compliance and find new opportunities to satisfy your customers, maybe by expanding your services to accost reported gaps, improving the quality of reported problem areas, or past elevating the client feel delivered by your contact center agents.

Contact Lens for Amazon Connect provides call transcriptions with rich analytics capabilities that can provide these kinds of insights, only y'all may not currently be using Amazon Connect. Yous need a solution that works with your existing contact center phone call recordings.

Amazon Machine Learning (ML) services like Amazon Transcribe Phone call Analytics and Amazon Comprehend provide characteristic-rich APIs that you tin use to transcribe and extract insights from your contact middle sound recordings at calibration. Although you could build your own custom call analytics solution using these services, that requires time and resource. In this post, we introduce our new sample solution for post call analytics.

Solution overview

Our new sample solution, Post Call Analytics (PCA), does most of the heavy lifting associated with providing an cease-to-end solution that tin process call recordings from your existing contact heart. PCA provides actionable insights to spot emerging trends, place agent coaching opportunities, and assess the general sentiment of calls.

Y'all provide your call recordings, and PCA automatically processes them using Transcribe Phone call Analytics and other AWS services to extract valuable intelligence such every bit customer and agent sentiment, call drivers, entities discussed, and conversation characteristics such as non-talk time, interruptions, loudness, and talk speed. Transcribe Call Analytics detects issues using built-in ML models that accept been trained using thousands of hours of conversations. With the automated call categorization adequacy, you can also tag conversations based on keywords or phrases, sentiment, and non-talk time. And you tin optionally redact sensitive customer information such equally names, addresses, credit carte du jour numbers, and social security numbers from both transcript and audio files.

PCA'south web user interface has a home page showing all your calls, as shown in the post-obit screenshot.

You can choose a record to see the details of the phone call, such as speech characteristics.

Y'all tin can also scroll down to come across annotated turn-by-turn phone call details.

You tin search for calls based on dates, entities, or sentiment characteristics.

You can also search your phone call transcriptions.

Lastly, you can query detailed phone call analytics data from your preferred business intelligence (BI) tool.

PCA currently supports the following features:

Transcription
- Batch turn-by-turn transcription with support for Amazon Transcribe custom vocabulary for accurateness of domain-specific terminology
- Personally identifiable information (PII) redaction from transcripts and audio files, and vocabulary filtering for masking custom words and phrases
- Multiple languages and automatic linguistic communication detection
- Standard audio file formats
- Caller and agent speaker labels using channel identification or speaker diarization
Analytics
- Caller and amanuensis sentiment details and trends
- Talk and non-talk time for both caller and agent
- Configurable Transcribe Call Analytics categories based on the presence or absence of keywords or phrases, sentiment, and non-talk time
- Detects callers' main issues, activeness items, and outcomes using built-in telephone call summarization ML models in Transcribe Telephone call Analytics
- Discovers entities referenced in the telephone call using Amazon Embrace standard or custom entity detection models, or simple configurable string matching
- Detects when caller and agent interrupt each other
- Speaker loudness
Search
- Search on call attributes such as time range, sentiment, or entities
- Search transcriptions
Other
- Detects metadata from audio file names, such as telephone call GUID, agent's name, and call date time
- Scales automatically to handle variable call volumes
- Bulk loads large athenaeum of older recordings while maintaining chapters to process new recordings as they arrive
- Sample recordings and so you can rapidly try out PCA for yourself
- It's easy to install with a single AWS CloudFormation template

This is just the first! We await to add many more exciting features over fourth dimension, based on your feedback.

Deploy the CloudFormation stack

Start your PCA feel by using AWS CloudFormation to deploy the solution with sample recordings loaded.

Utilize the following Launch Stack button to deploy the PCA solution in the us-east-i (N. Virginia) AWS Region.

The source code is available in our GitHub repository. Follow the directions in the README to deploy PCA to additional Regions supported by Amazon Transcribe.

For Stack proper noun, employ the default value, PostCallAnalytics.
For AdminUsername, utilise the default value, admin.
For AdminEmail, apply a valid email address—your temporary password is emailed to this address during the deployment.
For loadSampleAudioFiles, change the value to truthful.
For EnableTranscriptKendraSearch, modify the value to Yes, create new Kendra Alphabetize (Developer Edition) .

If you accept previously used your Amazon Kendra Free Tier allowance, yous incur an hourly cost for this index (more information on cost later in this postal service). Amazon Kendra transcript search is an optional characteristic, so if yous don't need it and are concerned about price, use the default value of No.

For all other parameters, apply the default values.

If you desire to customize the settings later, for example to apply custom vocabulary to better accuracy, or to customize entity detection, you tin can update the stack to set these parameters.

Select the two acknowledgement boxes, and choose Create stack.

The main CloudFormation stack uses nested stacks to create the following resources in your AWS account:

Amazon Simple Storage Service (Amazon S3) buckets to hold build artifacts and call recordings
AWS Systems Managing director Parameter Store settings to store configuration settings
AWS Stride Functions workflows to orchestrate recording file processing
AWS Lambda functions to procedure audio files and turn-by-plough transcriptions and analytics
Amazon DynamoDB tables to store call metadata
Website components including S3 bucket, Amazon CloudFront distribution, and Amazon Cognito user puddle
Other miscellaneous supporting resources, including AWS Identity and Access Management (IAM) roles and policies (using least privilege all-time practices), Amazon Uncomplicated Queue Service (Amazon SQS) bulletin queues, and Amazon CloudWatch log groups.
Optionally, an Amazon Kendra alphabetize and AWS Amplify search application to provide intelligent call transcript search.

The stacks take about 20 minutes to deploy. The main stack status shows equally CREATE_COMPLETE when everything is deployed.

Set your password

After you deploy the stack, you need to open the PCA web user interface and set your password.

On the AWS CloudFormation console, choose the principal stack, PostCallAnalytics, and choose the Outputs tab.
Open your web browser to the URL shown every bit WebAppURL in the outputs.

Y'all're redirected to a login page.

Open up the email your received, at the email address you provided, with the subject "Welcome to the Amazon Transcribe Post Call Analytics (PCA) Solution!"

This email contains a generated temporary password that you can utilize to log in (as user admin) and create your own password.

Set a new password.

Your new password must take a length of at least eight characters, and comprise uppercase and lowercase characters, plus numbers and special characters.

You're at present logged in to PCA. Because y'all set loadSampleAudioFiles to truthful, your PCA deployment at present has iii sample calls pre-loaded for you to explore.

Optional: Open the transcription search web UI and set your permanent countersign

Follow these boosted steps to log in to the companion transcript search web app, which is deployed just when y'all set EnableTranscriptKendraSearch when yous launch the stack.

On the AWS CloudFormation console, choose the master stack, PostCallAnalytics, and choose the Outputs tab.
Open your web browser to the URL shown as TranscriptionMediaSearchFinderURL in the outputs.

Y'all're redirected to the login page.

Open up the electronic mail your received, at the email address you provided, with the subject "Welcome to Finder Web App."

This email contains a generated temporary password that you tin can use to log in (as user admin).

Create your own password, simply like you already did for the PCA web application.

As before, your new password must take a length of at to the lowest degree eight characters, and contain capital letter and lowercase characters, plus numbers and special characters.

Y'all're at present logged in to the transcript search Finder application. The sample audio files are indexed already, and fix for search.

Explore postal service call analytics features

Now, with PCA successfully installed, yous're ready to explore the call assay features.

Domicile page

To explore the home folio, open the PCA web UI using the URL shown every bit WebAppURL in the main stack outputs (bookmark this URL, you'll use it often!)

You already have six calls listed on the home page, sorted in descending time order (most contempo offset). These are the sample audio files.

The calls have the following key details:

Chore Name – Is assigned from the recording audio file name, and serves as a unique chore name for this call
Timestamp – Is parsed from the sound file proper name if possible, otherwise it's assigned the time when the recording is candy by PCA
Customer Sentiment and Client Sentiment Tendency – Show the overall caller sentiment and, importantly, whether the caller was more positive at the finish of the call than at the showtime
Language Code – Shows the specified language or the automatically detected dominant language of the call

Call details

Cull the nigh recently received phone call to open and explore the telephone call detail page. You can review the call data and analytics such as sentiment, talk time, interruptions, and loudness.

Scroll down to meet the post-obit details:

Entities grouped past entity type. Entities are detected by Amazon Comprehend and the sample entity recognizer cord map.
Categories detected by Transcribe Call Analytics. Past default, there are no categories; see Call categorization for more information.
Summary tabs provide succinct summaries of the important components in agent-customer calls, including problems, activity items, and outcomes. For more information, see Phone call summarization.

Scroll further to see the turn-by-turn transcription for the call, with annotations for speaker, time marker, sentiment, interruptions, issues, and entities.

Apply the embedded media player to play the telephone call sound from any point in the conversation. Set the position by choosing the time marking annotation on the transcript or past using the role player time control. The audio player remains visible as you lot scroll downwardly the page.

PII is redacted from both transcript and audio—redaction is enabled using the CloudFormation stack parameters.

Search based on call attributes

To attempt PCA'due south built-in search, cull Search at the top of the screen. Under Sentiment, choose Average, Client, and Negative to select the calls that had average negative customer sentiment.

Choose Clear to try a different filter. For Entities, enter Hyundai and so choose Search. Select the call from the search results and verify from the transcript that the customer was indeed calling virtually their Hyundai.

Search call transcripts

Transcript search is an experimental, optional, add-on feature powered by Amazon Kendra.

Open the transcript web UI using the URL shown equally TranscriptionMediaSearchFinderURL in the main stack outputs. To notice a recent telephone call, enter the search query customer hitting the wall.

The results show transcription extracts from relevant calls. Utilize the embedded audio role player to play the associated section of the telephone call recording.

Y'all can aggrandize Filter search results to refine the search results with boosted filters. Cull Open Call Analytics to open the PCA call detail page for this call.

Query call analytics using SQL

You can integrate PCA call analytics data into a reporting or BI tool such as Amazon QuickSight by using Amazon Athena SQL queries. To try it, open up the Athena query editor. For Database, choose pca .

Observe the tabular array parsedresults. This table contains all the turn-by-plow transcriptions and analysis for each call, using nested structures.

You tin can also review flattened event sets, which are simpler to integrate into your reporting or analytics awarding. Use the query editor to preview the data.

Processing flow overview

How did PCA transcribe and clarify your telephone call recordings? Let'south take a quick await at how information technology works.

The following diagram shows the main data processing components and how they fit together at a high level.

Call recording sound files are uploaded to the S3 bucket and folder, identified in the primary stack outputs as InputBucket and InputBucketPrefix, respectively. The sample call recordings are automatically uploaded because you fix the parameter loadSampleAudioFiles to true when you deployed PCA.

Every bit each recording file is added to the input bucket, an S3 Event Notification triggers a Lambda function that initiates a workflow in Step Functions to process the file. The workflow orchestrates the steps to first an Amazon Transcribe batch task and procedure the results by doing entity detection and additional preparation of the call analytics results. Processed results are stored as JSON files in another S3 bucket and folder, identified in the main stack outputs as OutputBucket and OutputBucketPrefix .

Equally the Step Functions workflow creates each JSON results file in the output bucket, an S3 Result Notification triggers a Lambda function, which loads selected call metadata into a DynamoDB table.

The PCA UI web app queries the DynamoDB table to retrieve the listing of processed calls to display on the home page. The call detail folio reads additional detailed transcription and analytics from the JSON results file for the selected call.

Amazon S3 Lifecycle policies delete recordings and JSON files from both input and output buckets after a configurable retentivity period, defined past the deployment parameter RetentionDays. S3 Event Notifications and Lambda functions keep the DynamoDB table synchronized as files are both created and deleted.

When the EnableTranscriptKendraSearch parameter is true, the Pace Functions workflow too adds time markers and metadata attributes to the transcription, which are loaded into an Amazon Kendra index. The transcription search web application is used to search call transcriptions. For more information on how this works, meet Make your audio and video files searchable using Amazon Transcribe and Amazon Kendra.

Monitoring and troubleshooting

AWS CloudFormation reports deployment failures and causes on the stack Events tab. Meet Troubleshooting CloudFormation for help with common deployment bug.

PCA provides runtime monitoring and logs for each component using CloudWatch:

Step Functions workflow – On the Step Functions console, open the workflow PostCallAnalyticsWorkflow. The Executions tab show the status of each workflow run. Choose any run to see details. Cull CloudWatch Logs from the Execution event history to examine logs for any Lambda office that was invoked by the workflow.
PCA server and UI Lambda functions – On the Lambda console, filter past PostCallAnalytics to see all the PCA-related Lambda functions. Choose your function, and choose the Monitor tab to see function metrics. Choose View logs in CloudWatch to inspect function logs.

Cost cess

For pricing information for the chief services used by PCA, run across the post-obit:

Amazon CloudFront Pricing
Amazon CloudWatch pricing
Amazon Cognito Pricing
Amazon Comprehend Pricing
Amazon DynamoDB pricing
Amazon API Gateway pricing
Amazon Kendra pricing (for the optional transcription search feature)
AWS Lambda Pricing
Amazon Transcribe Pricing
Amazon S3 pricing
AWS Step Functions Pricing

When transcription search is enabled, you lot incur an hourly cost for the Amazon Kendra alphabetize: $one.125/hr for the Programmer Edition (offset 750 hours are free), or $1.forty/hour for the Enterprise Edition (recommended for production workloads).

All other PCA costs are incurred based on usage, and are Complimentary Tier eligible. After the Free Tier assart is consumed, usage costs add upwardly to almost $0.fifteen for a v-minute call recording.

To explore PCA costs for yourself, use AWS Cost Explorer or cull Bill Details on the AWS Billing Dashboard to come across your month-to-date spend past service.

Integrate with your contact center

You can configure your contact center to enable call recording. If possible, configure recordings for two channels (stereo), with customer audio on i channel (for case, aqueduct 0) and the agent audio on the other channel (aqueduct ane).

Via the AWS Command Line Interface (AWS CLI) or SDK, copy your contact center recording files to the PCA input bucket binder, identified in the main stack outputs as InputBucket and InputBucketPrefix. Alternatively, if you already save your phone call recordings to Amazon S3, employ deployment parameters InputBucketName and InputBucketRawAudio to configure PCA to employ your existing S3 bucket and prefix, so you lot don't have to re-create the files again.

Customize your deployment

Use the following CloudFormation template parameters when creating or updating your stack to customize your PCA deployment:

To enable or disable the optional (experimental) transcription search characteristic, use EnableTranscriptKendraSearch.
To use your existing S3 bucket for incoming call recordings, use InputBucket and InputBucketPrefix.
To configure automated deletion of recordings and call analysis data when using auto-provisioned S3 input and output buckets, employ RetentionDays.
To observe call timestamp, agent proper name, or call identifier (GUID) from the recording file name, use FilenameDatetimeRegex, FilenameDatetimeFieldMap, FilenameGUIDRegex , and FilenameAgentRegex.
To use the standard Amazon Transcribe API instead of the default call analytics API, use TranscribeApiMode. PCA automatically reverts to the standard style API for audio recordings that aren't compatible with the call analytics API (for example, mono channel recordings). When using the standard API some call analytics, metrics such as issue detection and speaker loudness aren't available.
To set up the listing of supported sound languages, use TranscribeLanguages.
To mask unwanted words, use VocabFilterMode and prepare VocabFilterName to the proper name of a vocabulary filter that yous already created in Amazon Transcribe. See Vocabulary filtering for more than data.
To improve transcription accuracy for technical and domain specific acronyms and jargon, gear up VocabularyName to the proper noun of a custom vocabulary that you already created in Amazon Transcribe. Come across Custom vocabularies for more data.
To configure PCA to utilise single-channel audio by default, and to place speakers using speaker diarizaton rather than channel identification, use SpeakerSeparationType and MaxSpeakers. The default is to use channel identification with stereo files using Transcribe Telephone call Analytics APIs to generate the richest analytics and about accurate speaker labeling.
To redact PII from the transcriptions or from the audio, set CallRedactionTranscript or CallRedactionAudio to true. Run into Redaction for more than information.
To customize entity detection using Amazon Cover, or to provide your own CSV file to define entities, use the Entity detection parameters.

Meet the README on GitHub for more details on configuration options and operations for PCA.

PCA is an open-source projection. Y'all can fork the PCA GitHub repository, enhance the code, and send united states of america pull requests so nosotros tin incorporate and share your improvements!

Clean up

When you're finished experimenting with this solution, clean up your resource by opening the AWS CloudFormation console and deleting the PostCallAnalytics stacks that you deployed. This deletes resource that you created by deploying the solution. S3 buckets containing your sound recordings and analytics, and CloudWatch log groups are retained subsequently the stack is deleted to avert deleting your data.

Live Telephone call Analytics: Companion solution

Our companion solution, Live Call Analytics (LCA), offers existent time-transcription and analytics capabilities by using the Amazon Transcribe and Amazon Comprehend real-time APIs. Different PCA, which transcribes and analyzes recorded audio after the phone call has ended, LCA transcribes and analyzes your calls equally they are happening and provides existent-time updates to supervisors and agents. You can configure LCA to shop call recordings to the PCA's ingestion S3 bucket, and use the two solutions together to get the all-time of both worlds. Come across Live call analytics for your contact center with Amazon language AI services for more information.

Decision

The Mail Call Analytics solution offers a scalable, toll-effective approach to provide call analytics with features to help better your callers' feel. It uses Amazon ML services like Transcribe Call Analytics and Amazon Cover to transcribe and extract rich insights from your customer conversations.

The sample PCA application is provided every bit open source—use it equally a starting signal for your ain solution, and help usa make it better by contributing back fixes and features via GitHub pull requests. For expert assistance, AWS Professional Services and other AWS Partners are here to help.

We'd dearest to hear from y'all. Allow u.s.a. know what you think in the comments section, or use the problems forum in the PCA GitHub repository.

About the Authors

Dr. Andrew Kane is an AWS Principal WW Tech Lead (AI Language Services) based out of London. He focuses on the AWS Language and Vision AI services, helping our customers architect multiple AI services into a unmarried use-instance driven solution. Earlier joining AWS at the showtime of 2015, Andrew spent two decades working in the fields of signal processing, financial payments systems, weapons tracking, and editorial and publishing systems. He is a keen karate enthusiast (just i belt away from Blackness Belt) and is besides an avid dwelling-brewer, using automated brewing hardware and other IoT sensors.

Bob Strahan is a Principal Solutions Architect in the AWS Language AI Services team.

Connor Kirkpatrick is an AWS Solutions Engineer based in the United kingdom of great britain and northern ireland. Connor works with the AWS Solution Architects to create standardised tools, code samples, demonstrations, and quickstarts. He is an enthusiastic rower, wobbly cyclist, and occasional bakery.

Franco Rezabek is an AWS Solutions Engineer based in London, UK. Franco works with AWS Solution Architects to create standardized tools, code samples, demonstrations, and quick starts.

Steve Engledow is a Solutions Engineer working with internal and external AWS customers to build reusable solutions to common problems.

shupespones.blogspot.com

Source: https://aws.amazon.com/blogs/machine-learning/post-call-analytics-for-your-contact-center-with-amazon-language-ai-services/