The customer wanted to enhance the response on the Virtual Assistant on Amazon Connect by implementing Artificial Intelligence capabilities on AWS Lex APIs for Natural Language Processing & Natural Language Understanding. This would help the system replace the current directed speech flow, towards an open-ended speech resolution format for the customers.
The identified use cases for the implementation are listed below:
- User Authentication
- Policy Fulfilment
- Policy Modification
- Policy Cancellation
- File a Claim
- Claim Status
- Policy Coverage
CSA Travel Protection is dedicated to providing customized travel insurance, related products, and comprehensive administrator services to their business partners and the traveling public, through quality customer service, innovative systems, and analytical expertise. With over 20 years of experience, CSA has developed a reputation for standing behind its customers and offering products and services to meet their needs.
About the Customer
CSA Travel Protection proudly delivers quality travel insurance and emergency services to their business partners and customers. Founded in 1991 and based in San Diego, California, CSA combines customer-driven innovation and risk management to create strategic solutions for our customers and business partners.
The customer support team receives 5000+ calls on a daily basis, while 80% of these calls were services related & post-sales support. Most of these queries being repetitive & standardized in nature, the client was looking to automate these queries on their call center, so the customer care team can look into more critical queries.
A solution which can integrate with Genesys for a seamless handoff to the automated system
Amazon Web Services provides on-demand cloud computing platforms and APIs which are reliable, scalable on a metered pay-as-you-go basis. It provides the necessary features which exactly fits the customer’s requirement to run their business.
The customer has applications running in their on-premise Data Centres and they decided to go with AWS since it provides low-cost migration services of the existing infrastructure to the AWS Cloud.
Along with that, AWS provides resources to aid in expansion and utilize more of their services for a long term basis with an appropriate Savings plan or reserving the capacity.
AWS is much more secure than a company hosting its own website or storage. It currently has dozens of data centers across the globe which are continuously monitored and strictly maintained. The diversification of the data centers ensures that a disaster striking one region doesn’t cause permanent data loss worldwide.
Why the Customer Chose the Partner
PowerupCloud is a cloud technology company with products & services in cloud migrations, managed cloud, and AI solutions. A Premier Consulting Partner of AWS and a Gold Partner of Microsoft Azure, Powerup provides product-led consulting solutions to enterprise customers on the public cloud platforms.
PowerupCloud has gained immense popularity over the last few years and successfully executed over 150 projects in cloud transformation. The customer was convinced that Powerup would be the best option who could deliver the services and fulfill their business requirement. It stood as a market leader in the Cloud Migration sector and managed to set up a strong customer base through its successfully executed projects. This encouraged the Customer to move ahead with PowerupCloud for assisting in the Cloud Migration Program and managing its services both in Cloud and on-premise.
What we proposed:
- The solution architecture scope is to enhance the response of the Virtual Assistant by implementing artificial intelligence capabilities for natural language processing and natural language understanding.
- Customer/user calls customer services line for policy details
- AWS Connect captures the user query and passes it to AWS Lex. Lex performs natural language processing.
- AWS Lex classifies user query to intent. It then looks for any entities the user might have referred to in the query. A virtual assistant prompts the user to provide details like policy number, email, zip code.
- It then passes the extracted parameters (intent and entities) to AWS Lambda function for call to action.
- Lambda calls DynamoDB for fetching the policy information using the extracted entities, which might include policy number, zip code, etc.
- It then completes the action based on the identified intent like policy modification.
How AWS Services were used as a part of this solution:
Automated flows for 7 use cases, which included user authentication using an alphanumeric policy #, filing a claim, claim refund, refund status & so on AWS Connect, which used AWS Lex for NLP classification & user query understanding
AWS Connect integrated with Genesys dial-in nos on the existing call center support system, with a seamless handover onto the voice automated system
The complete design was on a serverless architecture, with policy manipulation logic written on Lambda Functions on AWS
System integrated with live policy database via REST-based APIs for live policy updates & reading the latest policy information
Completely voice-based interaction, system handed off to a human agent if it is not able to resolve user query
Architecture Diagram and System Architecture
- Customer calls into the dedicated Direct Inward Dialling (DID) number and Total Frequency Deviation (TFD) numbers on Genesys
- Customer selects an option from the IVR to be redirected to the virtual assistant on Connect
- Customer inbound calls are passed to Amazon Connect through Genesys.
- Customer interactions are created using a contact flow engine. A contact flow is created as per the flow chart diagram for each of the use cases mentioned in the Conversational Flow.
- Amazon Lex understands speech and natural language. Lex maps the spoken words to corresponding actionable intents. Each intent performs an action in response to the human input.
- Any data access, retrieval, and updates are processed using Amazon Lambda.
- Data is passed through by calling Generali Global Assistance endpoints (GGA endpoints). The GGA endpoints are the REST APIs. GGA endpoints are used to fetch and validate all flows on policies.
- All the policies are in the TIPs database. The TIPs database is the truth data for policies. For all scoping and design purposes, TIPs will be out of the scope of VA
- DynamoDB and elastic search indices are kept in sync using create and update APIs.
- The DynamoDB policy table uses Optimistic Locking to handle concurrent update requests.
- Swagger UI details the supported endpoints.
35% reduction in call volumes to the agents
90% reduction in resolution time for the customers.
- Amazon Connect allows you to tap the very rich library of AWS Services including AI, Natural Language Processing, text to speech, transcriptions and translations.
- Amazon Connect instance is scalable, resilient, and fully redundant.
- Amazon Connect enables even a non-technical professional the simplicity of configuring an inbound call flow, attaching a phone number, defining agents and operating hours, and in less than an hour, taking phone calls.
About the Partner
Powerup is an enterprise-class cloud consulting company focused on helping businesses move to the cloud. The dynamic and intuitive team at Powerup strives for customer satisfaction and delivers desired outcomes to businesses with undivided attention and support. Our consulting team can help you implement, migrate, and optimize your cloud applications. We can also help you build Big Data analytics platforms to enable you to extract more value from your environment.
Founded in 2015, Powerup is a born-in-the-cloud, Premier Consulting Partner of AWS. It is the only AWS certified partner in APAC to have both Data & Analytics Competency and Machine Learning Competency, and one of the few certified partners of AWS with competencies across Migration, DevOps, Financial Services, Well-Architected, and Security & Compliance.