AWS adds machine learning capabilities to Amazon Connect

In a bid to help enterprises offer better customer service and experience, Amazon Web Services (AWS) on Tuesday, at its annual re:Invent conference, said that it was adding new machine learning capabilities to its cloud-based contact center service, Amazon Connect.

AWS launched Amazon Connect in 2017 in an effort to offer a low-cost, high-value alternative to traditional customer service software suites.

As part of the announcement, the company said that it was making the forecasting, capacity planning, scheduling and Contact Lens feature of Amazon Connect generally available while introducing two new features in preview.

Forecasting, capacity planning and scheduling now available

The forecasting, capacity planning and scheduling features, which were announced in March and have been in preview until now, are geared toward helping enterprises predict contact center demand, plan staffing, and schedule agents as required.

In order to forecast demand, Amazon Connect uses machine learning models to analyze and predict contact volume and average handle time based on historical data, the company said, adding that the forecasts include predictions for inbound calls, transfer calls, and callback contacts in both voice and chat channels.

These forecasts are then combined with planning scenarios and metrics such as occupancy, daily attrition, and full-time equivalent (FTE) hours per week to help with staffing, the company said, adding that the capacity planning feature helps predict the number of agents required to meet service level targets for a certain period of time.

Amazon Connect uses the forecasts generated from historical data and combines them with metrics or inputs such as shift profiles and staffing groups to create schedules that match an enterprise’s requirements.

The schedules created can be edited or reviewed if needed and once the schedules are published, Amazon Connect notifies the agent and the supervisor that a new schedule has been made available.

Additionally, the scheduling feature now supports intraday agent request management which helps track time off or overtime for agents.

A machine learning model at the back end that drives scheduling can make real-time adjustments in context of the rules input by an enterprise, AWS said, adding that enterprises can take advantage of the new features by enabling them at the Amazon Connect Console.

After they have been activated via the Console, the capabilities can be accessed via the Amazon Connect Analytics and Optimization module within Connect.

The forecasting, capacity planning, and scheduling features are available initially across US East (North Virginia), US West (Oregon), Asia Pacific (Sydney), and Europe (London) Regions.

The Contact Lens service, which was added to Amazon Connect to analyze conversations in real time using natural language processing (NLP) and speech-to-text analytics, has been made generally available.

The capability to do analysis has been extended to text messages from Amazon Connect Chat, AWS said.

“Contact Lens’ conversational analytics for chat helps you understand customer sentiment, redact sensitive customer information, and monitor agent compliance with company guidelines to improve agent performance and customer experience,” the company said in a statement.

Another feature within Contact Lens, dubbed contact search, will allow enterprises to search for chats based on specific keywords, customer sentiment score, contact categories, and other chat-specific analytics such as agent response time, the company said, adding that Lens will also offer a chat summarization feature.

This feature, according to the company, uses machine learning to classify, and highlight key parts of the customer’s conversation, such as issue, outcome, or action item.

New features allow for agent evaluation

AWS also said that it was adding two new capabilities—evaluating agents and recreating contact center workflow—to Amazon Connect, in preview. Using Contact Lens for Amazon Connect, enterprises will be able to create agent performance evaluation forms, the company said, adding that the service is now in preview and available across regions including  US East (North Virginia), US West (Oregon), Asia Pacific (Sydney), and Europe (London).

New evaluation criteria, such as agents’ adherence to scripts and compliance, can be added to the review forms, AWS said, adding that machine-learning based scoring can be activated.

The machine learning scoring will use the same underlying technology used by Contact Lens to analyze conversations.

Additionally, AWS said that it was giving enterprises the chance to create new workflows for agents who use the Amazon Connect Agent Workspace to do daily tasks.

“You can now also use Amazon Connect’s no-code, drag-and-drop interface to create custom workflows and step-by-step guides for your agents,” the company said in a statement.

Amazon Connect uses a pay-for-what-you-use model, and no upfront payments or long-term commitments are required to sign up for the service.