While conversational artificial intelligence (AI) and automation are ideal for optimizing the entire conversation, it’s important to realize that the agent’s effort isn’t over when the interaction ends. What happens after the call can be just as important to your contact center’s business outcomes as what happens during the conversation.
That’s because the time spent in after-call work (ACW)—including categorizing and summarizing the call, updating systems, and taking follow-up actions on promises made to the customer during the interaction—impacts average handle time, call waiting times, customer satisfaction (CSAT), costs, and agent productivity and satisfaction.
The longer it takes agents to complete their ACW tasks, the longer other customers must wait to be helped, which can result in additional costs to the business. At the same time, when agents are too rushed or stressed when addressing ACW, accuracy can suffer and follow-up tasks may be missed, leading to customer disappointment and repeat calls, which affect first call resolution (FCR) rates. By using conversational AI and robotic process automation (RPA) to automatically handle ACW and promises management, contact centers can achieve a remarkable return on investment by improving agent productivity and accuracy, shortening average handle and call waiting times, increasing customer satisfaction, and reducing costs due to fewer callbacks on the same issue.
As clear as these benefits are, understanding how to get started and choosing the right conversational AI and automation solution can be a daunting task. Contact center leaders are challenged to see beyond the buzzwords and hype to the practical, yet sophisticated advances in conversational AI that can help them transform their operations with automation.
This guide can help. You’ll learn about the business case, the essential capabilities you’ll need (explained in layperson terms), the right questions to ask, and how to choose a partner to help you achieve a rapid, sustainable return on investment.
What you’ll learn in this guide
The business case for automating ACW
Capabilities required to automate ACW
What to look for in a conversational AI and automation solution
Questions to ask vendors as you evaluate solutions
When it makes sense to build it yourself
Developing your business case for automating ACW
While most companies begin their conversational AI and automation journey by automating self-service, you can also start by automating ACW. Either way, whether you’ve already automated one or more use cases or are just getting started, formulating your business case for automating ACW should be the next step.
A good place to start is asking questions to understand what your agents need to accomplish today as part of their workflow once the interaction has ended. Which tasks are your agents asked to perform after each call?
While the individual tasks vary by industry, ACW typically includes:
Researching and recalling what transpired during the call
Categorizing and logging the call
Documenting / summarizing the interaction in call notes
Updating information in various systems such as customer relationship management (CRM) software
Performing follow-up tasks such as sending confirmation emails, triggering fulfillment, and other actions
Once you understand the work output, systems, and tasks involved, it’s time to look at your baseline ACW metrics, including:
How much time do our agents spend on average on ACW?
How much of their time on average is spent handling call summarization as part of ACW?
How long are wait times for customers before they can speak with an agent?
How accurate are call notations and summaries?
How are commitments (also known as promises) logged today?
How are commitments fulfilled?
How many repeat calls are due to missed commitments and follow-ups?
What additional time is added to call handle times to address repeat and escalated calls due to missed commitments?
With this information, you can start assessing the benefits for your company of automating all or part of the ACW effort.
Here are some of the ways that automation can help.
ACW Task: Create call summary
Automation
During the interaction with the customer, a conversational AI and automation solution automatically listens and transcribes the call in real time. After the call ends, the call summary is automatically created and presented to the agent to edit and confirm.
Benefits
- Reduces the duration of call summary work from an average of 2.5 minutes to 30 seconds
- Improves quality and accuracy
- Creates a better customer experience as agents handling future calls can accurately understand what happened in the past
- Improves agent productivity
- Reduces call waiting time by freeing up agents to take the next call
ACW Task: Update systems
Automation
A conversational AI and automation solution automatically updates the CRM system as well as others and provides automated call disposition to maintain the quality of call categories.
Benefits
- Reduces ACW and average handle time
- Improves call categorization accuracy for trend analysis
- Improves agent productivity and reduces call waiting time
ACW Task: Manage commitments
Automation
With automated promises management, conversational AI and RPA extract and track promises and commitments made by agents in real time, align expectations with customers immediately following the call, and manage fulfillment after the call
Benefits
- Reduces repeat calls and improves customer satisfaction
- Shortens average handle time (AHT) by reducing ACW time and effort
- Improves agent productivity
- Reduces call waiting time by freeing up agents
- Reduces the number of screens agents need to log into to execute workflows that register and fulfill promises made
Intelligent Applications
A $19 Million Business Case
A major telecommunications company automated ACW, including call summaries and categorization, for its 16,000 agents. The automation is projected to save the company $19 million annually based on the reduction in agent effort.
Understanding the capabilities needed for ACW automation
While basic natural language processing (NLP) is frequently used to automate transcription tasks, creating true business value requires much more than a simple transcription of the call. Achieving the level of automation and accuracy needed to create the business outcomes described earlier demands a combination of advanced technologies centered around conversational AI.
A conversational service automation (CSA) platform combines the following sophisticated technologies to optimize and automate the entire customer conversation and journey:
Conversational AI
A set of advanced AI technologies that recognizes and comprehends human language in multiple languages and uses this understanding to optimize and analyze conversations in and across multiple channels.
Intelligent Applications
AI-powered software called intelligent applications includes rules engines, user interfaces, notifications, and alerts, and other components that handle specific use cases within the contact center, such as intelligent agent assistance, intelligent self-service, and others.
Natural Language Processing and Understanding (NLP/NLU)
NLP and NLU are components of conversational AI that help computers understand and interpret human language.
AI Analytics/Intelligent Decision Support
AI-powered software called intelligent applications includes rules engines, user interfaces, notifications, and alerts, and other components that handle specific use cases within the contact center, such as intelligent agent assistance, intelligent self-service, and others.
Reporting and Dashboards
Built-in reporting provides realtime insight into conversational performance and the customer journey.
Video Sensor and Services
As the number of video interactions increase, conversational AI combined with facial cognition, facial gesture understanding, demographic identification, and gaze enable engagement and emotional comprehension of the interaction.
Robotic Process Automation (RPA)
RPA is software that can emulate the actions of a human interacting with digital systems to automate repetitive tasks and end-to-end business processes.
Voice Biomatrics and Security
This capability applies advanced conversational AI to agent verification, using ongoing voiceprints to authenticate agents continuously during their shifts.
All of these core technologies work together in a CSA platform to create an operational backbone that enables contact centers to quickly and effectively deploy, scale, and expand conversational AI-powered automation. Automating ACW is one of the use cases for which a CSA platform is ideal.
Intelligent Applications
- Intelligent Self-service
- Intelligent Agent-Assistance
- Intelligent Agent-Authentication
- Intelligent Decision Support
Automation Services
- Conversational AI
- Workflow
- RPA
Mutli-lingual AI Services
- Speech Sensor & Services
- NLP Sensor & Services
- Interaction Sectioning
- Video Sensor & Services
- Authentication & Screen Services
- Conversation & Journey Insights
Evaluating solutions for ACW automation
To automate and optimize ACW for your contact center, you need to choose a platform built for enterprise-scale automation. While superficial AI solutions and open source software can provide sandbox environments for exploring AI, they don’t provide the sophisticated capabilities seen in leading-edge conversational AI technology designed with the contact center in mind.
While many products may claim to provide conversational AI, NLP, and RPA capable of being used to automate ACW, it pays to look deeper at the specific capabilities this use case requires. Before choosing a solution, look for a platform that offers:
Industry-leading speech recognition
Your platform should understand your agents and customers with a high degree of accuracy. To do so, the technology must be able to listen for and detect the language, including specific dialects.
Look for a solution with the highest voice accuracy and broadest language coverage possible to achieve maximum accuracy for ACW automation.
Intent recognition
Beyond understanding the language, the conversational AI and automation solution must be able to identify and understand customer intent, parsing the conversation in real time to accurately categorize the call and understand how to act on the assurances made by the agent during the call.
Insist on a solution that offers continuous tuning and optimization to deliver the highest degree of accuracy possible in predicting true intent. Higher accuracy reduces agent effort and improves the customer experience as well as agent performance.
Intelligent interaction sectioning
Intent recognition alone is not enough to optimize the entire conversation. The solution must also be able to automatically identify and section the key classifying elements in a conversation and match them to categories to add context and facts to an intent. For example, the platform needs to automatically identify the greeting, discover key issues and intents during the call, and recognize and record the outcome/resolution with an accurate interaction sectioning capability. With a holistic view of the interaction, the platform can accurately summarize the call, update appropriate systems, and understand assurances made as well as record the outcomes.
Ask whether the solution you’re considering relies on manual training and inputs from humans to drive interaction sectioning capabilities. Leading solutions will use machine learning to guide the technology in recognizing and classifying elements. Make sure the platform or solution you’re considering can automatically identify the interaction including greeting, discovery, and resolution accurately.
Customer sentiment and emotion recognition and analysis
Beyond understanding the context and intent of the conversation, it’s important to glean information about the customer’s feelings. First, there’s recognition and extraction of customer sentiment as positive, neutral, or negative. Sentiment provides insight into the customer’s state of mind to enhance empathy and communications. It also provides data that can be analyzed to uncover trends and patterns. Going a step further, understanding customer emotions such as frustration, disappointment, and anger helps the platform inspire customer loyalty and trust.
Look for a solution that offers not only sentiment recognition and analysis but emotion recognition and analysis as well to understand your customers’ feelings and take relevant actions that demonstrate empathy.
Promises management
A promise made that is not kept or tasks that are not performed correctly can quickly negate the positive effects of a good conversational experience. Promises management is an important component of ACW that directly impacts your call handling times, wait times, and customer satisfaction.
Choose a platform that offers a promises management application built using conversational AI and RPA to extract, track, and fulfill commitments made during agent and customer conversations.
Integrations with existing systems and architecture
The last thing your contact center needs is another siloed solution. You also don’t want to be on the hook for building custom integrations. That’s why it’s important to choose a platform that works with your existing architecture and interfaces to your current contact center solutions.
Ask about integrations with common contact center technologies, including call platforms, CRM systems, and other systems before you choose a platform.
Cloud-based services
As more companies are moving to the cloud for its agility, scalability, ease of management, and availability of new technologies, it’s important to choose a conversational AI and automation platform that is designed for use in the cloud.
Look for a solution that can take advantage of native cloud elasticity to scale AI modeling. Also ask whether the vendor supports high availability and other cloud services such as tenant management to seamlessly manage and create multiple tenants, an API gateway to enable connection to data, and insights and sensors from external systems. Finally make sure the solution is built on a scalable, enterprise-grade architecture.
Top three questions to ask before you buy
How quickly can your conversational AI and automation platform be implemented?
Can your conversational AI and automation platform be extended easily to support more AI-powered automation use cases?
When will we see return on investment with your solution?
Deciding whether to buy or build
With the availability of conversational AI toolkits, RPA development platforms, and open source NLP/NLU software, your company may be considering a do-it-yourself (DIY) approach to automating and optimizing the customer conversation.
While appealing in terms of upfront costs and project control, ultimately many companies underestimate the long-term cost and resource commitments of developing and maintaining their own conversational AI and automation capabilities over multiple years. Optimizing outcomes using conversational AI also requires deep expertise that many companies don’t have in house.
That said, if your company has the resources available internally to write, test, and maintain custom code, you may be able to create a custom solution in-house or with the help of consultants. However, before you decide, consider the following:
Scope: Are you automating one isolated part of the customer conversation or the entire journey? Which capabilities must be included? Which ones will you add in later iterations of the software?
Timing: How long will the project take? How long are you willing to wait before the automation is ready to use and can start delivering benefits?
Resources: How many engineering resources and AI/NLP experts will you need? Will you need to hire consultants?
Integration: How many systems will your solution need to integrate with? Do you have the resources to commit to maintaining custom integrations as contact center systems evolve?
Budget: Based on scope, resources, and integrations, what will be the total cost of the project?
Maintenance: What will it cost to scale and maintain the solution as new capabilities are introduced or new use cases supported?
Unless your company wants to own the intellectual property for your custom software and applications and sell it to other contact centers, choosing an off-the-shelf conversational AI and automation platform is almost always the faster, more cost-effective, and most beneficial option.
Choosing a Conversational AI and Automation vendor
Beyond evaluating the technology and capabilities you need to automate ACW and
other use cases, you also want to choose the right vendor. The ideal conversational AI and automation partner should offer:
- Visionary leadership and a technology roadmap that aligns with the vision
- Deep expertise in AI, NLP/NLU, RPA, and related technologies
- Domain expertise in optimizing customer experience/contact center operations
- Deployment methodology and services to help you achieve rapid time to value
- Security and privacy of customer data
Next steps
Automating ACW can have a dramatic impact on your costs, staffing, wait times, call quality, and customer satisfaction. Choosing the right platform to begin or further your automation journey starts with a good business case, understanding the capabilities you will need, and evaluating the options
