Amazon Connect initially disrupted with a pay-per-use price model but is now flexing its artificial intelligence muscles.
This week the largest cloud computing vendor, AWS, is holding its 11th re:Invent conference. Given its leading position in Infrastructure as a Service (IaaS), much of the news revolved around infrastructure.
Over the past few years, AWS has been offering a handful of SaaS services, one of which is its Connect contact center product, and the company announced several new capabilities to this solution.
AWS Connect Brings AI and ML with a Pay per Use Model
Amazon Connect, which is built on the same technology that Amazon customer service agents use worldwide, utilizes artificial intelligence (AI) and machine learning (ML) to deliver a personalized customer experience (CX). It’s a single solution with a simple user interface (UI) that companies can implement across multiple channels to connect with customers.
One of the concerns regarding cloud contact centers is that they don’t scale past a few thousand agents. This has never been the case with AWS as its own contact center is almost 100,000 agents. One of the unique attributes of Connect is that it uses a true “pay per use” model. With traditional CCaaS services, customers pay $X/month/user. With Connect, a business can provisioning as many agents as they need but only pay for utilization.
According to AWS, there are hundreds of thousands of customers making more than 10 million contact center interactions every day on Amazon Connect. Some of its big-name customers include Adidas, Adobe, and National Australia Bank, all of which have successfully rolled out interactive voice response (IVR). In the contact center, IVR provides callers with a self-service option before being connected to a live agent.
It’s particularly valuable for seasonal companies where call flows tend to rise and fall. I recently talked with the Global Head of Reservations for a major hotel chain and she said the pay-per-use model saved the company a significant amount of money throughout the pandemic as travel halted. As travel picked up, the company could add to its deployment without having to guess at how many agents to provision.
Agent Workspace Reduces Training Time
While the unique financial model has created the opportunity for AWS, its differentiation comes in AI-based advanced features.
Last year, AWS launched Amazon Connect agent workspace, which unifies all the tools agents need in one place to manage customer inquiries, view information, resolve issues, and handle calls, chats, tasks, and cases. At re:Invent, AWS introduced step-by-step guides for Amazon Connect Agent Workspace to reduce agent training time and resolve customer issues faster. These one-click actions are based on different triggers from customers, such as their profile information, case history, or past interaction from IVR.
The triggers help provide agents with the most relevant, guided actions that they can take to resolve an issue. For instance, if a customer is calling about a lost order, the agent workspace will display the customer profile, the billing details, shipping status of the order, and actions in the step-by-step guide for replacing or refunding the item.
Amazon Connect Flows Brings No-Code to IVRs
Supervisors can design guided experiences for agents using Amazon Connect flows, a no-code workflow tool for IVR. The guides are use-case dependent and customizable. There are relevant drag-and-drop-components that support the guides, which can be configured within Amazon Connect flows.
In addition to rolling out step-by-step guides for agent workspace, which is available now in preview, AWS improved the search and filtering function in flows as well.
“This is a great example of a capability where we heard consistent problems like inefficient apps and long, hard training for agents—all of which finds its way to end customers,” said Ryan Braastad, senior product marketing manager for Amazon Connect. “We had the ability to test this with Amazon customer service, and working closely with them, we were inspired to make this available as part of Connect.”
Machine Learning Improves Forecasting
The second announcement is around the simplicity and single-click functionality of Amazon Connect. Earlier this year, AWS launched Connect forecasting, capacity planning, and scheduling in preview. It allows companies to predict, allocate, and verify that they have the right number of agents scheduled at the right time to serve their needs in the contact center. This ML-powered staffing capability became generally available this week.
“The folks that do this for a living don’t have a lot of time. It’s usually a supervisor who also does the forecasting or planning or scheduling, and they want to get back to what they consider their main job as opposed to doing the administrative work,” said Tom Johnston, principal product manager for Amazon Connect.
One Amazon Connect customer, Litigation Practice Group, was running into challenges with forecasting based on the historical data the company had with its previous system. Once the law firm switched to Connect, it saw accuracy of over 95 percent for forecasts and schedules. For the firm, the simplicity, flexibility, and the high quality of data “were all main reasons why they liked using it,” said Johnston.
Lastly, AWS announced that Contact Lens for Amazon Connect will expand to include omnichannel support. AWS also added evaluation forms in Contact Lens to improve agent performance. Contact Lens is a tool that analyzes conversations between customers and agents through speech transcription, natural language processing (NLP), and sentiment analysis. While it’s been used primarily for voice, it can now analyze conversations across voice and chat channels in one place.
