How To Improve Customer Sentiment with AI

By Machielle Thomas

When customers are frustrated, they want a response fast. When they don’t get that speedy response, your customer sentiment falls.

83% of customers agree they feel more loyal to brands that respond to and resolve their complaints. And 12% of Americans rate their number one service frustration as “lack of speed.” 

The speed at which your customer service team can accurately resolve issues plays a massive role in sentiment—AI can help you get faster.

What does AI have to do with Customer Sentiment?

If good customer service can lead to positive sentiment, improving customer service is one really impactful way to improve sentiment.

These days, AI is at the forefront of improving customer service. It’s managing frustrated customers faster and more accurately than ever by interacting with them just like a human would.

Customers interact with automated agents, typically via chat or phone, and receive the same personalized experience a human would provide.

The personalized experience that only one-on-one support can provide is very important to customers—76% expect it.

Using AI, you can provide it at scale without expanding your team. 

4 Ways to Improve Customer Sentiment with AI

The fastest route to improving customer sentiment with AI is working with a technology partner. Start by sourcing someone who can talk to your customers like a person. You want a partner who:

  • Truly has the AI capability to sound like a human
  • Will treat your data and your customer’s data with care
  • Is really good at synthesizing large amounts of information

Forethought AI is the most powerful agent for customer support because the conversations we deliver are natural. They improve resolution rates and lower company costs.

Once you’ve found a technology partner that fits your needs, you’re four steps away from better sentiment: set up your tools, power your chatbot, route your support tickets, and make AI your agent’s personal assistant.

1. Set Your AI Tools Up for Success

AI is only as good as the information you give it. To train it, you want a strong, up-to-date knowledge base and a bank of past interactions with customer support.

You’ll want to connect your automated agents with your customer relationship management (CRM) tool for the most accurate responses. This will provide the context your automated agent needs to answer questions quickly or deliver complex solutions to trickier problems.

2. Let AI Power Your Chatbot

If you want every customer to get an immediate response and feel positive about your brand afterward, AI needs to be your first line of defense. There’s nothing faster or more accurate.

The easiest way to do this is to route all customer service requests through a chatbot powered by an automated agent. That way, AI takes the first pass at every customer question. 

If you’ve done a good job training it, AI can solve all but the most complex customer inquiries, which require relationship building. 

Take iFIT, a global health and fitness subscription technology company. They wanted to achieve excellence in customer support so that their customers would feel supported and love their brand. But they weren’t prepared to manually build chatbot workflows and guess what each customer wanted. 

So, they chose an AI solution that would learn as it went—Forethought Solve. Now, customers can ask questions about product details, cost, technical information, and more, like the one below about Nordic Track, one of their treadmills.

Today, 33% of chats are deflected by Forethought’s Solve widget. Their customers receive the information they need, and agents aren’t bogged down with repetitive questions.

“Something I really like about Forethought is that other companies think they do AI, but they’re really AI-driven chatbots. They still make you create workflows from start to finish. With Forethought, the AI knows where the customer needs to go. It makes it quicker, easier, and clearer for customers to engage with us. You feel the difference with their AI–it’s been huge for me.”

– Dustin Auman, Operations Manager

3. Let AI Route Your Support Tickets

Slow support can cause poor customer sentiment, which can be exacerbated by tickets waiting to be routed—even if an AI chatbot is used on the front lines. Imagine you’ve got a complex customer query that a person should solve. So, AI sends the ticket to the support team. 

It’s now sitting in a queue. The clock is ticking.

You’re still working with a team that manually reviews tickets to route them to the right place. It’s 10:00 pm local time, though, and they won’t be back until the morning. Their queue is full, so it might take them until lunchtime to even get this ticket routed.

Your customer may wait an entire day for the answer to their question. They may be frustrated by the time you respond, and when this happens at scale, customer sentiment can plummet.

AI can automatically route tickets to the right place around the clock. It’s smart and can understand the nuances of a request based on customer sentiment, location, industry, product, and more to route it to the right person.

Take UpWork, for example, a freelance marketplace that connects independent contractors and employers worldwide. They get support requests from both sides—and lots of them.

They use Triage to route those requests to the right place.

Triage helps the team instantly understand the sentiment of each ticket. It detects patterns in text to automatically classify the overall emotion of the inquiry as positive, negative, or neutral. 

By understanding which types of tickets have the most negative sentiment, the team redirects resources to create better content and workflows to address those topics. Today, they achieve 90% accuracy and coverage in classifying new tickets.

4. Let AI Assist Your Support Team

AI can support your agents just as well as your customers. Imagine an agent live with a customer, over chat, or on the phone. Those agents also need to get answers to questions quickly. 

Typically, they pull information from a knowledge base or an internal wiki. They have a ton of tabs open and are trying to find the answer to complex problems as fast as possible while a customer is on hold.

Maybe they’ll even tap another customer service rep on the shoulder for help, putting the customer on hold for longer or escalate a problem to their boss if they can’t figure it out.

AI can also help by surfacing the correct information for the agents to close the ticket as quickly as possible and send the customer on their merry way with renewed positive sentiment.

Uncommon Goods’ team uses Assist to assist its agents. 

They’re an online catalog retailer selling unique, authentic, and well-crafted products and experiences. They get a lot of support tickets, especially around the holidays, when they onboard hundreds of new agents with little experience answering customer questions.

Assist surfaces relevant knowledge articles, macros, past tickets, and notes directly within their agent’s helpdesk, Zendesk, and into non-helpdesk forms like email. 

When a customer submits a ticket, Assist provides their agent with a relevant help center article to attach to the case. It also applies macros—templated responses that live in help desks—with placeholder texts for the agent. Then, agents personalize the message as they see fit before sending it to the customer.

This helps new agents service customers fast, without too much lag time as they get up to speed, and without ruining customer sentiment toward Uncommon Goods.

How to track customer service AI’s impact on sentiment

You need some sort of natural language processing (NLP) to analyze sentiment because you need to be able to tell if the words customers use are positive, negative, or somewhere in between.

Typically, sentiment analyses cover comments on social media platforms or review sites. Still, you can do the same in customer service interactions to determine whether customers are happy with your brand.

It’s also useful to ensure that tickets get routed to the right place so that customers are happy with customer service in the first place and to track what customers like or don’t like about your products.

Track Emotions to Provide Better Support Experiences

AI-powered chatbots can assess customer sentiment during conversations and tailor responses accordingly by classifying customer responses as positive, neutral, or negative.

This helps make sure tickets get routed to the right place so that they’re appropriately escalated.

It’ll also help ensure tickets go to someone who speaks the customer’s native language and gives agents the proper context before they enter a conversation.

Kickfin provides digital tipping software to restaurants. They use Triage to determine the sentiment of the support inquiries, empowering agents with personalized classifiers to understand them at scale. 

When reviewing case history, Kickfin’s support leadership team easily understands which customers are happiest and which are most unhappy overall with accurate sentiment tags. 

Kickfin has been able to predict customer sentiment with 96% accuracy, allowing for confidence in decision-making when evaluating customer happiness.

Track Sentiment for Specific Products or Services

You can take it a step further and analyze the sentiment to determine if customers like a new product or feature or learn which parts of your product are causing the most confusion or problems.

Take a mobile app, for example. Customers chatting with automated agents give it positive comments about its user-friendly interface but negative comments about slow loading times. 

AI can flag these issues, clearly defining where the focus should be. Engineering teams can then make interface changes and prioritize updates directly impacting the customer experience.

Who can benefit the most from customer service AI?

We recently conducted a research project that analyzes how AI impacts the customer experience. In that report, we analyzed who was happiest with their AI tools.

We found that Jira and Freshdesk customers were happiest with their AI tools.

We also found that companies in the education, and minerals & mining spaces were happiest with how AI helped their customer experience.

The report is packed with many more benchmarks on topics like the types of solutions, the average number of tickets resolved by industry, the revenue other businesses generate when they implement AI, and more.

Download Forethought’s AI in CX Benchmark Report 2024

The First Step is Choosing a Technology Partner

Improving your customer sentiment means keeping your customers happy, and a big part of that is a good experience when they talk to customer service.

Customers want the quickest way to accurate, helpful solutions. Forethought’s AI-powered agents can help you do that at scale.

Request a demo today to talk with an expert about how AI might improve your customer sentiment.

Dive in.

Interested in generative AI for customer support? Check out this guide to learn about the 3 key pillars you need to get started.

Get your copy
Decor Half Circle Orange
Call to action decor Call to action decor