Thanks for submitting the form.
What is Amazon Connect?
Amazon Connect is an omnichannel cloud contact centre. One can set up a contact centre easily using amazon connects user-friendly platform in a few steps. Also, one can personalize the end customers' experience by utilizing omnichannel communication channels. For instance, both chat and voice channels can be offered to customers simultaneously, reducing wait times and increasing customer satisfaction.
Amazon Connect is an open platform that means you can integrate it with other applications such as salesforce according to your requirements. One can use the services provided within the ecosystem of aws such as AI and ML services.
A serverless interactive query service or interactive data analysis tool which is used for processing complex queries and in a lesser amount of time. Click to explore about, Amazon Athena Architecture
Services Support with AWS connect
The above image gives the various aws services that can be used with aws connect. Let’s see the details of AI and ML services.
Amazon Connect uses the following services for ML/AI:
- Amazon Lex— Let's create a chatbot as Interactive Voice Response (IVR).
- Amazon Polly— Provides text-to-speech in all contact flows.
- Amazon Transcribe— Capture conversation logs from Amazon S3 and converts them to text for you to review.
- Amazon Comprehend—Perform transcriptions of recordings and apply machine learning speech analytics to calls to identify opinions keywords, comply with company policies, and more.
Solutions Using AWS Connect
Business Problem: Real-time customer insights For contact Centre using machine learning in Multi-channel communication(voice and chat).
The given solution offers two salient features: the first is Real-time analysis of the customer conversation, and the other is Multi-channel communication. Below are the details of each feature.
This feature provides Machine learning capabilities to the contact centre by integrating amazon connect and aws contact lenses for amazon connect(service). This enables businesses to analyze the customer conversation for sentiment, trends, and compliance conversation in real-time to make quick and intelligent decisions. By real-time, we mean one can analyze the conversation while it is happening in the contact centre. This is one of the features of this Solution.
Intelligent Chat communication channel
The other feature is to provide an intelligent chat communication channel to the end customer by integrating amazon’s lex service; this enables a contact centre to leverage the power of NLU (Natural language Understanding), the same technology that powers amazon’s Alexa bot. This will empower the contact centre to maximize customer satisfaction through both channels. More details about the use cases of these features are given in the Feature section.
A type of data warehouse service in the Cloud which is fully managed, reliable, scalable and fast and is a part of Amazon’s Cloud Computing. Click to explore about, Guide to Amazon Redshift
In the architecture diagram,
- Starting with the user, the first building block is AWS connect. Here we will design the contact flows for our contact centre. This means the overall functioning of the contact centre will be defined here. This will include:
- Defining Agent Profiles
- Defining queue (section like sales, technical support)
- All administration policies will be defined here.
- From Aws connect block further, two flow emerges one is for enabling real-time analysis and the other for enabling Intelligent chat support, defining both of them in detail:
1- The high-level steps included in real-time Analysis Flow:
- First, we Set up Amazon Connect with a contact flow that enables Contact Lens real-time capabilities.
- The Real-time analysis metrics such as sentiment score are generated by aws contact lens in aws connect block.
- The Analysis metrics are sent using AWS lambda and aws gateway API to the agent ccp page.
- Agents can access these real-time metrics based on them. Intelligent decisions can be made at the agent’s end.
- Admin will always be able to see overall real-time metrics. Also, they can control what type of information is transferred to the Agent.
2- The high-level steps for setting up Amazon Lex chat service
- Amazon Connect receives an incoming call and initiates an Amazon Connect contact flow. This Amazon Connect stream captures the caller's speech and transmits it to Amazon Lex.
- Amazon Lex starts the requested bot. The Amazon Lex bot translates the caller's speech into text and sends it to AWS Lambda via an event.
- AWS Lambda accepts incoming data, transforms or enhances it as needed, and calls external APIs via several transport methods
- The external API processes the content sent to it from AWS Lambda.
- The external API returns a response to AWS Lambda.
- AWS Lambda accepts a response from an external API and then forwards it to Amazon Lex.
- Amazon Lex returns the response to Amazon Connect.
- Amazon Connect processes the response.
AWS SageMaker uses Jupyter Notebook and Python with boto to connect with the s3 bucket, or it has its high-level Python API for model building. Click to explore about, Amazon SageMaker
What are the features of the solution proposed AWS Connect?
The features of the solution proposed AWS Connect are below mentioned:
- Detailed analytics and sentiment analysis
- With the aws contact lens enabled in the aws connect contact flow, you can use Machine learning, natural language processing, and speech-to-text analytics to discover essential customer insights. For example, one can perform the following tasks easily.
+ Analyze call transcripts
+ Analyze sentiments
+ Detecting trends in customer conversations.
+ Detects issues by analyzing transcripts.
- Advanced conversational search
- This solution also provides accessibility to a fast full-text search while on calls, i.e. in real-time. One can search by desired keywords, customer or agent sentiment scores, and by “non-talk” time by applying easy GUI-based filters on the dashboard.
This analysis provides the common utterances during the calls that end with positive or negative sentiment, allowing the supervisors to increase customer satisfaction.
- Real-time alerts
- One can create flags for any customer experience issues considered critical by the contact centre; the alerts will be generated based on matching keywords from the conversation and the defined categories from the contact centre admin.
For instance, if the customer says “cancel my subscription”, this can generate real-time alerts so that the contact centre team can make intelligent decisions.
- Real-time call transcripts
- Real-time call transcripts can also be obtained for each call. This will be helpful in a scenario where:
+ The call is transferred from one agent to another, which will help the routed agent get the context easily without repeating the same procedure.
- Automating Repetitive Task with Lex integration
- The Lex integration can automate several repetitive tasks for end customers like booking an appointment and cancelling orders without human agents' interference. This will have the following benefits.
+ Waiting time for human agents will be reduced.
+ Less human agent resources will be required.
What are the business values of AWS Connect?
Detailed analytics and sentiment analysis
|- Quantify the perception of customers in real-time.
- Enable intelligent decisions based on the sentiments.
Advanced conversational search
|- Provide a quick glance to businesses about what happy and less satisfied customers are uttering to make important decisions for increasing the satisfaction of customers.|
|- Help businesses detect crucial customer-related issues in real-time.|
Real-time call transcripts
|- Real-Time insights from ongoing calls in form of transcripts|
Automating Repetitive Task with Lex integration
|- Reduction of waiting time for Agent.
- Multichannel accessibility for customers
- Automation of repetitive tasks.
Features availability as compared with Observe AI
AI and Analytics Capabilities for Real-time Analysis
Real-time metrics for Agents can be extracted out to built similar feature
Agent Coaching and Feedback
Agent feedback along with agent metrics can be used to develop a similar feature
We have covered the High-level architecture for developing AI, and ML-powered contact centres using aws connect and essential use cases of the solution. The design will lay the foundation for developing the solution.