Just think for a moment about the number of times you’ve called a customer service hotline in the last year. How does that compare to five years ago? Ten years ago?
Table of contents
- What Are AI and NLP?
- 6 Ways AI and NLP are Transforming Customer Service
- AI complements, rather than replaces, human skills
- Setting up a better solution for NLP for customer support
- Industry use cases for NLP contact center tools
If you’re like most customers, you rely more and more on alternate channels like text messaging, web chat, and social media when you need to get in touch with a company you do business with.
With access to a wealth of customer data, technology at their fingertips, and a massive workforce with which to test and implement innovative customer service solutions, the contact center industry is a prime candidate for the deployment of leading-edge tools like artificial intelligence (AI) and natural language processing (NLP).
What Are AI and NLP?
AI is the branch of computer science focused on leveraging data to enable machines to perform the types of tasks that typically rely on human intelligence.
Between 2018 and 2019, the portion of companies currently using AI in their operations grew from 4% to 14%. Gartner predicts that by 2021, 15% of all customer service interactions will be handled entirely by AI.
NLP is a branch of AI that focuses specifically on distilling meaning from human language, be it spoken or written. We’re beginning to see it pop up in more and more applications in our everyday lives.
Virtual assistants like Alexa and Siri use NLP to process their users’ commands. Online language translation tools use it to ensure consistent meaning between the input and the output. Google’s autocomplete feature uses it to predict your most likely intention based on the first few words of your search query.
And now, contact centers are using NLP to understand precisely what a customer is calling about so the best solution can be reached as quickly as possible.
6 Ways AI and NLP are Transforming Customer Service
Added context enables greater personalization
AI is a game-changer for its ability to bring a new level of context to call center interactions. With NLP, we can combine sentiment analysis (is the caller’s tone pleasant? Strained? Angry?) with historical data (what was the question they typed into a webchat two days ago?) to anticipate their needs before they even have to say a word to an agent.
As you can imagine, this facilitates faster interactions that lead to speedier, more efficient resolutions. But it also adds a layer of personalization that customers have come to expect. With content gleaned from AI, agents can have a more personalized conversation than if they were starting from a blank slate with every caller.
Smart call routing saves time and labor
Ai-powered virtual agents and interactive voice (IVR) systems are great for helping manage call volume and giving customers an option to self-resolve simple issues. When a live agent is needed, however, AI serves a different purpose. It can be leveraged to route each call directly to the agent who’s best equipped to handle it using data on the customer’s inquiry, the agent’s expertise, how long they’ve been waiting on hold, and more.
Repetitive tasks can be automated
One of the biggest personnel challenges for call centers is keeping agents engaged. The repetitive nature of the work can hurt morale and worse, contribute to performance issues and turnover.
AI can eliminate many repetitive tasks that have traditionally been done manually, like searching for a piece of customer information. Instead, with AI, the required data point is immediately pulled up onto the screen. This allows agents to prioritize work that actually uses their customer service skills rather than wasting time clicking and scrolling, which leads to higher engagement.
Pattern discovery makes for better use of resources
If we drill down to the basics, AI’s core function is to uncover patterns found within thousands or even millions of data points that would otherwise be impossible to discern. Call centers can harness this capability to translate patterns into meaningful information that can drive business decisions.
For example, by using AI to process historic call center data, you might learn that customers who make contact on the weekend are more likely to be calling about issue X, while those on weekday mornings are typically experiencing issue Y. Armed with such information, you can make better decisions about which agents to schedule during those times based on their specific skills.
Customers can communicate in their native language
One of the most immediately useful applications of NLP is real-time language translation, where phone calls made or text messages sent in different languages can be automatically translated into the language of the agent. This facilitates communications across international borders while increasing customer satisfaction and reducing costs that might otherwise be spent on outsourcing translation duties.
This NLP application also presents an amazing opportunity to better serve people with disabilities or speech impairments and even deaf people who use sign language.
Supervisors can easily access the most relevant data
A call center manager could easily spend all day poring over activity or performance reports, but who has the time for that? With AI solutions, we can prompt our call center software to feed supervisors only the most important metrics, prioritizing alerts about the most pressing issues.
This allows supervisors to spend more time focusing on big-picture tasks that increase efficiency and reduce costs rather than getting bogged down in the minutia of individual interactions.
AI complements, rather than replaces, human skills
We don’t anticipate AI and NLP will replace human agents any time soon. After all, customers still find great value in the option to speak with a live agent when they need to. Rather, AI’s greatest promise lies in its ability to complement the skills of human agents.
With AI call analysis, for example, agents might receive live coaching alerts that help them be more effective in the customer interaction, like prompts on which question to ask next or scripts for how to alleviate a customer’s concern. With the help of NLP, the agent can tailor their responses to the customer’s mood, like upselling if the caller sounds excited or diffusing tension if they’re expressing frustration.
The new frontier of AI in the contact center combines technology with human talent in a way that makes each of them greater than the sum of their parts. It presents an exciting opportunity to harness data in a way that helps staff do their jobs faster, more effectively and with more meaning.
Setting up a better solution for NLP for customer support
Create a communication flow
When establishing a speech analytics program, many organizations fail to develop effective communications channels between the speech analytics team and the rest of the business. This can be extremely disruptive.
Successful businesses set up open communication flows so that other departments have visibility on the program and know where and how to obtain information or results. It’s equally important to avoid conflict by developing a robust request process with agreed-upon SLAs to avoid differing expectations around what, how, and how quickly the speech analytics team can provide deliverables.
Create a scalable, repeatable process
Developing a continuous improvement cycle is important to ensure that you continue to build on the improvements that speech analytics allows you to make and that your program doesn’t stagnate. A good practice would be to implement the DMAIC data-driven quality improvement process from the Six Sigma methodology.
- Define the problem or use case in terms of the target to accomplish, e.g., Reducing Average Handle Time (AHT).
- Measure and gather all relevant data about the use case (i.e., Silence %, Current AHT).
- Analyze the data that you have gathered and work to determine the root causes that are affecting performance in your chosen area of focus.
Determine the problems speech analytics will solve
Without proper planning, poor execution is sure to follow. Creating a strategic vision for speech analytics adoption based on clear intended use cases, KPIs, and enterprise opportunities will provide the focus needed to accomplish desired outcomes and create a culture around the tool.
Don’t try to boil the ocean. Prioritize one or two key initiatives to start with and communicate the significance of these metrics to all stakeholders. Use the experience gained during these first pursuits to begin building robust processes.
Industry use cases for NLP contact center tools
The CX improvements and customer satisfaction possible with speech analytics are vast, but below are a few examples of use cases by industry.
Call center agents have to document important patient information while they’re on a live call. This information is incredibly important for additional context around the interaction and informs future patient care so it should be accessible to providers for analytics purposes alongside the audio and transcribed conversation.
Speech analytics captures call audio and rich metadata, and uses sophisticated automatic speech recognition engines through natural language processing, to create rich, text-based files to categorize patient conversations for AI engines to reason over both in realtime and for post-call analysis.
Speech analytics and natural language processing solutions can be applied in financial services firms to raise the bar on personalization capabilities and meet ever-changing compliance demands.
Call recording, regular call calibrations, and keyword flagging allow financial services organizations to define the characteristics of an ideal customer interaction and then monitor and vet each customer interaction or individual sales or marketing campaign accordingly and set the right evaluation criteria for agents so CX goals are always met.
Lead generation centers can use speech analytics to track keywords and phrases during prospect interactions Themes in your calls can drive transformational insights for your lead pipelines and help reveal the words that are mentioned most often by prospects. Your team can then review the transcripts to learn how to optimize lead nurture flows. Sales representatives can also use those same insights for generating campaign and messaging ideas and strategic planning.