Amazon Connect turned five years old this spring. And, while Amazon Web Services (AWS) doesn’t have a long history as a contact center vendor, there are plenty of reasons that make Amazon Connect a viable Contact Center as a Service (CCaaS) offering, such as having more than a decade of experience running high-scale customer support with about 100,000 agents.

Omdia view

Summary

Amazon Connect turned five years old this spring and, while Amazon Web Services (AWS) doesn’t have a long history as a contact center vendor, there are plenty of reasons that make Amazon Connect a viable Contact Center as a Service (CCaaS) offering, such as having more than a decade of experience running high-scale customer support with about 100,000 agents.

Amazon Connect: What’s past is prologue

Amazon Connect recently celebrated its fifth anniversary, which started with the availability of its cloud-native contact center offering in the spring of 2017. While Amazon doesn’t have a long history as a contact center vendor, the company isn’t exactly a novice when it comes to the technology. More than 12 years ago, the company struggled to find a contact center solution that could support personal interactions and the need to scale up and down to accommodate the retailer’s cyclical business and extraordinary growth. So, Amazon built its own contact center solution. After customers saw, first hand, how Amazon was able to automate personal interactions at scale, they asked the company to make its homegrown system commercially available. And Amazon Connect was born. Today, with tens of thousands of AWS customers supporting more than 10 million contact center interactions per day on Amazon Connect, it is one of the fastest growing services in the history of AWS. And there’s good reason for this growth.

For starters, Amazon Connect offers omnichannel capabilities for voice, chat, and messaging in one app, so agents don’t have to jump from window to window to support customers along their journeys. This benefit cannot be understated, as streamlining the number of tools agents use decreases handle times, which improves agent and customer satisfaction.

Amazon Connect takes a similar approach to its broader offering by consolidating and unifying contact center tools into a single application, so centers don’t have to buy and stitch together different components. This also means that agents, managers, and supervisors get one application for an agent interface, agent management, a workflow builder, an interactive voice response (IVR) system, automatic call distributor (ACD) functionality, routing, and analytics. The offering also enables centers to improve customer self-service experiences with the use of automation via IVR and chatbots that leverage natural language speech and text capabilities.

To help keep costs down, all of this is offered through Amazon Connect’s pay-as-you-go pricing model, which provides the flexibility to scale the service up or down, as needed. Inbound and outbound voice usage has a per-minute fee of $0.018, while high-volume outbound voice usage costs $0.025 per minute.

A look under the hood

Some key components of Amazon Connect’s offering include Voice ID, Customer Profiles, Contact Lens, and Wisdom.

Amazon aims to streamline the customer authentication process with its Amazon Connect Voice ID. With customers’ permission, they can be enrolled in Voice ID, which captures their speech and creates an enrollment voiceprint. The next time they call, Voice ID can authenticate them when interacting with the IVR system or an agent by leveraging its real-time and passive voiceprint authentication capability. Voice ID then creates an authentication confidence score from 0 to 100, with 90 being the default threshold. The agents then receive real-time results, “authenticated” or “not authenticated,” enabling them to complete the authentication process or take additional steps to do so. Voice ID also enables centers to create a watchlist with audio recordings of known fraudsters, making it easier to identify bad actors and prevent fraudulent attacks.

Amazon Connect Customer Profiles automatically brings customer data (name, home address, telephone number, email address, purchase history, etc.) from multiple sources to the agent or IVR system. The data is brought into the Amazon Connect agent application, so they don’t have to tab between different windows. Data can be culled from the Amazon Connect contact history or from Salesforce, Zendesk, ServiceNow, Marketo, Shopify, Segment, or homegrown apps through APIs. One enhancement made to Customer Profiles since its launch is the ability to leverage machine learning to find and remove duplicate records across disparate systems.

Amazon Connect Contact Lens gives managers better visibility into customer sentiment, trends, and compliance violations. It can alert managers or supervisors when customer sentiment drops in real time to offer the agent guidance via chat or to transfer the call. It can also be used for post-call analysis. At Amazon’s re:Invent conference in late November last year, the company unveiled a machine-learning call summarization capability for Connect. This enables Contact Lens to automatically summarize important parts of customer conversations (i.e., issues, actions, and outcomes). These key parts can also be expanded to view the full transcript. Managers can view the call summary to quickly understand the context of customer interactions, so they don’t have to waste time wading through entire call transcripts. If they find something worth investigating, they can click through to view the full transcript.

Finally, Amazon Connect Wisdom offers agent assistance capabilities, including machine-learning-powered search and real-time recommendations. Wisdom was made generally available in September 2021, and enables agents to type questions or phrases into the Wisdom search box without knowing the most appropriate keyword to type. Wisdom understands what the agent is searching for and returns the results to agents in their preferred agent application. Wisdom can collect information from third-party applications, such as Salesforce, ServiceNow, and Zendesk. Additionally, to provide agent assistance, Wisdom can use Contact Lens real-time analytics to detect customer issues, find related content in the connected repositories, and provide recommendations to help agents resolve the issue.

In addition to these main components, the company has been investing heavily in machine learning, enabling Amazon Connect to offer some advanced capabilities, such as the following:

  • Intelligent conversational chatbots and IVR experiences with Amazon Lex and Polly
  • Real-time caller authentication that uses machine learning to verify customer identity with Amazon Connect Voice ID
  • Automation of identity resolution with Amazon Connect Customer Profiles
  • Real-time agent assistance that enables agents to quickly access the information they need to solve issues in real time, using Amazon Connect Wisdom
  • Real-time speech analytics to analyze the sentiment of agent/customer calls with Contact Lens for Amazon Connect
  • Omnichannel enablement through voice chat and messaging.

In addition to these features, clients can integrate with third-party applications using Amazon Connect’s open platform.

Some considerations

Before investing in Amazon Connect, though, there are a few considerations contact center leaders must contemplate.

First, any vendor that offers pay-as-you-go pricing for inbound and outbound voice and messaging presents this as a benefit to enterprise customers, suggesting that it is a more flexible and efficient pricing model than a flat fee. It sounds fair to only pay for what you use. However, because contact centers experience daily, monthly, or occasional large and sometimes unplanned call spikes, the costs for this model can be unpredictable. This doesn’t mean organizations should avoid this payment model, simply that they should plan based on historical interaction volumes and leave some wriggle room for surprises.

Amazon is also previewing what it’s calling a “high-volume outbound communication” capability. When the offering is generally available, it will enable organizations to utilize a predictive dialer that can reach millions of customers per day by automating these outbound calls and using artificial intelligence to notify agents when there is a real voice on the other end.

While some organizations abuse this type of capability by inundating recipients with “spray-and-pray” telephone marketing campaigns, there are some more customer-friendly use cases. For example, large healthcare networks can send appointment confirmations via text and email and automatically follow up with a call to patients who don’t respond. Additionally, airlines could use this capability to reach a lot of customers suddenly affected by flight schedule changes.

Conclusion

Amazon’s technology strengths lie in its fresh perspective, simple approach, and open and cloud-native architecture. Because of these capabilities, Amazon Connect enables organizations to think differently about contact center technologies. No longer should organizations be saddled by big, complicated contact center technologies that are difficult to maintain and upgrade. Amazon’s cloud-native solution aims to bring more flexibility to contact centers with quick cloud-based installation times (often measured in days and hours), easy scaling, painless upgrade options, and flexible pricing. When you combine these capabilities with Amazon’s brand recognition, huge customer base, and deep pockets, it makes Amazon Connect a viable omnichannel, cloud-native contact center platform for the future.

Appendix

Further reading

Event Recap: Enterprise Connect 2022 (April 2022)

Assessing the Disruptive Impact of Amazon and Twilio on the CX Market (March 2019)

 

Author

David Myron, Principal Analyst, Business Platforms and Applications

askananalyst@omdia.com