Amazon Web Services Inc.’s cloud contact center service Amazon Connect is getting some major under-the-hood updates today aimed at improving the experience for both call agents and customers.

Amazon Connect is an omnichannel cloud contact center service launched in 2017 that businesses can use to enable customer service and customer engagement. Customer service representatives can respond to phone calls or chat inquiries from customers in the same manner as if the contact center infrastructure were set up and managed on-premises.

With the launch of Amazon Connect Cases in preview today, agents have access to a new case management feature that makes it easier for them to track customer cases, collaborate with other staff and resolve their issues faster.

As Amazon explains, case management is one of the main tasks of customer service operations. Customer problems around issues such as payments and billing, product defects and service changes often require multiple conversations with different people before they are resolved. In such cases, customer service agents use case management tools to track these issues.

However, such tools add a lot of complexity to the contact center infrastructure stack. Moreover, case management tools are not always that helpful, as they mean agents are forced to waste time switching between multiple applications to look up details.

Amazon Connect Cases aims to resolve this by automatically creating a new case the first time a customer calls. It then tracks all related calls, chats and tasks and keeps them in one place, so agents have a quick understanding of the case at their fingertips.

Chatbots can also access the case information. So when a customer needs to talk or chat with a human agent, Amazon Connect Cases will automatically route them to the most suitable agent with the relevant case information attached. Amazon said that helps to improve the average handle time of each case and often leads to first-contact resolution.

Meanwhile, Amazon said, the introduction of Amazon Connect Outbound Campaigns gives companies a more cost-effective way to contract hundreds or even millions of customers each day. The service, generally available now, can be used to power marketing promotions, for appointment reminders or upcoming delivery notifications without the need to integrate any third-party tools. Contact center managers can schedule and launch high-volume outbound communications by specifying the communication channel, contact list and content to be delivered.

Amazon Connect Outbound Campaigns makes life easier with a predictive dialer that automatically calls customers on a list, with outreached throttled according to agent availability. It also uses a machine learning model that can discern if the person who picks up is a human or a voicemail greeting, helping increase agent efficiency by only connecting them to a live customer.

Accenture Plc Senior Technology Delivery Manager Shawn Hatton said his company helped to set up an Amazon Connect contact center for a state Medicaid agency client in order to manage incoming and outgoing health insurance enrollment calls. The agency required an easy way to contact thousands of people about their health insurance and leveraged Amazon Connect Outbound Campaigns to do this.

“Previously, each individual call center agent spent up to 17 hours per month manually dialing nearly 1000 different phone numbers,” Hatton said. “With outbound campaigns, the calls are placed automatically, increasing overall agent bandwidth and taking 80% less time to complete the same number of calls.”

A final update pertains to the general availability of the Amazon Lex Automated Chatbot Designer, which makes it easier for businesses to design and create chatbots trained on their own transcripts.

As Amazon explains, the process of training artificial intelligence-powered chatbots can take weeks. Chatbots are trained to identify the customer’s intent by listening to hours of recorded conversations and reading thousands of lines of chat transcripts.

The Amazon Lex automated chatbot designer can accelerate chatbot training times to a matter of hours instead of weeks, Amazon claims. Developers feed their transcripts into Lex, which can analyze thousands of lines in just hours and use them as the basis of an initial chatbot design.

The generated bot design can be focused on common intents, associated phrases, and be fed with a list of information that’s required to solve customer issues, such as the customer’s insurance policy number and the claim type. Once this foundation is built, developers can iterate on the design, adapting various chatbot prompts and responses until they get it just right.

Photo: guvo59/Pixabay