Architecting Cross-channel Intelligent Customer Engagements

Recently, we have had customers express the desire to build “omni-channels.” These omni-channels provide a centralized overview of digital engagement channels that help you better understand your customers and offer a more personalized experience.

Many companies have tried or are trying to implement an omni-channel strategy. However, because most existing channels are built on different platforms and by different vendors, they do not always integrate with each other easily. Consequently, most businesses end up spending most of their time figuring out integration and compatibility before they can extract customer insights.

By using various AWS services, you can build digital customer interaction interfaces, such as chatbot, call centers, and Alexa. You can also use traditional interfaces like email, SMS, and push notifications. You can then store and analyze the data in a data lake and offer personalized experiences using artificial intelligence (AI). This blog will walk you through how to use relevant AWS services, enabling you to focus on delivering what really matters to your customers.

Building an AI-enabled cross-channel platform on AWS

In the “Customer channels” section of Figure 1, the categories outlined in section 1 (email, SMS, notifications, chatbot, call center, voice skills, augmented reality [AR]/virtual reality [VR]) show what channels a business usually has. Section 2 shows how customers can choose to engage through website, applications, social media, phone call, voice assistant, or display. As shown in the figure, traditionally, there are email and SMS. As mobile applications are becoming the main digital interface, there has been a huge increase in using push notifications. In recent years, chatbot is becoming more popular, which enables customers to use mainstream social media platform messaging services. What’s more, speaking is still the most natural way of communication. Voice calls and the latest voice assistants, such as Amazon Echo, are used widely. AR/VR that contains rich media and offers immersive experience is also becoming a success.

Figure 1. Enterprise customer engagement channels and corresponding AWS services

Now let’s see which AWS services from Figure 1 can help you set up and manage these channels. The service names mentioned in the following list are followed by a number that corresponds to their placement in Figure 1.

• Amazon Simple Email Service (Amazon SES) (3) is a cost-effective, flexible, and scalable inbound/outbound email service. Amazon Simple Notification Service (Amazon SNS) (4) also provides a fully managed messaging service.
• For chatbot, Amazon Lex (5) provides a serverless service for building a conversational interface into any application using voice and text.
• With Amazon Connect (6), you can build a fully managed cloud-based contact center that offers a seamless experience across voice and chat in minutes.
• With Amazon Alexa Skills Kit (7), you can build Alexa skills on Amazon Echo.

In addition, after building an Amazon Lex chatbot, the same bot can be used in Amazon Connect and Alexa.

The next sections show you how to use AWS services to architect an intelligent cross-channel customer engagement platform to extract insights.

Build one channel and extend to multiple with Amazon Lex

You can build just one Amazon Lex chatbot and reuse it across different channels, including social media, mobile, call center, Alexa, and even AR/VR. With this strategy, you avoid creating your chat engines repetitively on different channels, and your customers will have seamless experiences when they switch channels. What’s more, you can focus on managing the channels and improve customer interactions by shifting the undifferentiated heavy-lifting of system integrations between these channels to AWS. Here is a video demo that will show you how to do it.

Personalize channels with Amazon Pinpoint and Amazon Personalize

As shown in Figure 2, most of the AWS services mentioned so far fall under Customer Engagement on AWS, which Amazon Pinpoint is the anchor service of. Amazon Pinpoint is a flexible and scalable multi-channel marketing communication service. It enables you to engage customers by sending targeted messages via multiple channels, such as text, email, voice, mobile push, and custom channels through API operations.

Figure 2. AWS customer engagement services with Amazon Pinpoint

To create a more personalized user experience, you can use Amazon Personalize, a fully managed Machine Learning Service on AWS (Figure 3). You can start by providing data to Amazon Personalize, such as activities, items, and users. In a few clicks, you get a custom model trained and hosted for you and start offering recommendations through a private API. After that, you can integrate your personalization models into Amazon Pinpoint via the console or API operations. To incorporate Amazon Pinpoint into your website or mobile apps, refer to the detailed guide provided in the Predictive User Engagement using Amazon Pinpoint and Amazon Personalize blog post.

Figure 3. AI-powered cross-channel customer engagement with Amazon Personalize

Extract customer insights with Analytics and Machine Learning on AWS

Figure 4 shows you how to expand the architecture and dive deeper to understand your customers. The data analytics and AI/ML sections of Figure 4 show services that can help you gain more insights from your customer data and interactions. (Many of the service names mentioned in the following two paragraphs are followed by a number that corresponds to their placement in Figure 4.)

Figure 4. Intelligent cross-channel customer engagement with Analytics on AWS and Amazon AI/ML services

With Analytics on AWS, you can combine various customer data sources and analyze them to get a more comprehensive understanding of your customers. For example, with AWS Lake Formation (1) you can build a secure data lake in days. You can use AWS Glue (2) for preparing and loading data. Amazon Athena (3), a serverless interactive query service, can analyze data in Amazon Simple Storage Service (Amazon S3) using standard SQL. Then you can use Amazon Redshift (4) for data warehousing, Amazon EMR (5) for big data processing, and Amazon QuickSight (6) for data visualization.

With Machine Learning on AWS, you can even further leverage these datasets. For example, you can apply Amazon Transcribe (7) to convert your customers’ speech files to text quickly. By combining Amazon Transcribe with Amazon Connect, you can get your customer calls automatically transcribed to create a fully searchable archive. This feature has already been built by AWS and is called Contact Lens for Amazon Connect. If your customer prefers a different language, Amazon Translate (8) allows you to localize content for international users and easily translate large volumes of text efficiently. You can also use Amazon Textract (9) to instantly read and process any types of documents, extracting text, forms, and tables without manual effort. And there is also Amazon Comprehend (10) to help you uncover the insights and relationships in your unstructured data. All these AI services are pre-trained by AWS to offer you ready-made intelligence, no machine learning (ML) skills are required. If you want a customized ML model to suit your specific business need, you can use Amazon SageMaker (11). SageMaker is a fully managed service that provides you with the ability to build, train, and deploy your custom models quickly.

Sample use cases to get started

Our solution showcases how to integrate various channels by using multiple AWS services and products. You can treat them as individual building blocks and don’t need to implement all services in one day. Here are some example use cases to get started:

• If you want to create a chatbot and meanwhile upgrade your call center, you can consider using Amazon Lex and Amazon Connect.
• If you want to create an email campaign with personalized attributes, you can start with Amazon Pinpoint and Amazon Personalize.
• If you want to get more insights from your unstructured and semi-structure data like emails, support tickets, product reviews, or even social media, you can use Amazon Comprehend, or even Amazon SageMaker to train custom ML models.
• If you are in the financial industry and want to provide a new digital banking service to engage your customers, you can use Amazon Alexa to create a newer, more modern banking experience.

Figure 5. Sample use cases to get started

Conclusion

To better enable you to interact with your customer across different channels, you can use Customer Engagement, Analytics, and Machine Learning on AWS to build a digital omni-channel platform with a real-time feedback loop. With that, your company can offer an AI-powered, personalized experience to your customers without spending significant time managing the underlying platforms and undifferentiated heavy lifting, you just focus on what really matters to you and your business.

Figure 6. Intelligent cross-channel customer engagement with real-time feedback loop

Related information

Watch “Architecting intelligent cross-channel customer engagements (Level 300 – Advanced)” at AWS Summit Online ASEAN on May 19. Register now.