If you could wave a magic wand and transform your customer support team’s workload and CX productivity, chances are, you’d want them focused on stuff that generates revenue.
That doesn’t usually include manually routing tickets, pulling reports, digging through internal wikis with customers on hold, or answering the same questions over and over again, like “What are your hours of operation?” These days, artificial intelligence (AI) can realistically take over these tasks for your team.
That leaves the humans you employ to focus on conversations with customers who need the most help and building relationships with high-value clients—the stuff that really impacts your bottom line.
We’ll walk through the details of how 10 real-world companies dramatically increased their customer support productivity with AI tools so you can copy their success.
1. Increasing CX Productivity by eliminating time spent on simple questions
As your business grows, so does the number of questions that your customers need answered. You end up with a live customer support team that repeatedly answers the same simple questions.
“How long do I have to return this?”
“How do I reset my password?”
“How do I download my invoices?”
“Do you have any discount codes?”
These eat up your team’s time and are a major CX productivity drain. Not to mention, your team’s probably not excited to face each day when they know they’re not going to get the chance to be creative and really connect with a customer because they’re battling troves of these simple questions in their queue.
They know if they let them sit, they’ll watch customer satisfaction scores (CSAT) decline and first response time (FRT) increase, which will probably get someone in management asking questions. So, they give the simple questions the same level of attention and care as the customers who enter the queue asking more complex questions like these:
“I started my return already and accidentally boxed up the wrong item before it shipped. Can you send it back to me so I can return the correct item?”
“I’m trying to integrate your SaaS product with Marketo, but the email field doesn’t look like it’s properly syncing. Everything else is working just fine, though. Can you help?”
“We just checked into our Airbnb, and it looks like the cleaning crew accidentally left materials here. Can you send someone to come get them?”
These customers really need to talk to someone who can help them with a unique, complex problem. Talking with a smart, helpful customer support person who can address their problem quickly will turn around their sentiment and hopefully keep them loyal spenders.
You have two options—hire more people so everyone gets addressed faster, or find a solution that takes the simple questions off your support team’s plate.
An AI-powered automated agent could pick up this Slack without expanding your headcount. Solve is designed to manage even the most nuanced questions through its natural language processing (NLP) and machine learning capabilities. It sounds human, gives the right information, and closes tickets without needing a live agent.
Example 1: A sports management platform handles a seasonal surge
Spordle is a sports management platform that connects governing bodies, leagues, clubs, associations, teams, parents, and fans within one integrated platform. Their ticket volume surges when it’s time to play ball (or not, depending on the sport).
One year, one of their biggest clients, Hockey Canada, made a few changes that significantly increased registrations. Customer support ticket volumes reached an all-time high for their small support team of three full-time agents and one part-time agent. They were overwhelmed by nearly 7,000 tickets per month.
The year prior, the ticket surge was so severe that eight additional employees had to take on support roles, working 14+ hour days just to keep up. This was unsustainable, and with registration numbers expected to grow even more, Spordle knew they had to do something.
They used Solve to automate responses to routine inquiries, like how to sign up for tryouts, which deflected over 21,000 tickets in just three months.
Solve was successful because Spordle allowed Forethought’s generative AI models to be trained on data from previous customer service interactions. That way, Solve understood details like sentence structure, meaning, and tone. Plus, there are no crappy decision trees. They used Solve’s workflow builder to detect customer intent and automatically build a workflow for each inquiry.
Example 2: Crypto platform replaces a chatbot that’s more work than it’s worth
Abra is a full-service crypto platform that struggled with its old chatbot, which was bundled with its help desk. It was clunky, unintuitive, and required many manual updates. Too many tickets still needed to be escalated to live agents, which overwhelmed them.
They replaced their old chatbot with an AI-powered automated agent using Solve. It can handle even the most nuanced questions without ever having to manually tinker with a workflow—those are automatic.
“The best part about using Forethought is their partnership with Abra to create a better customer experience. We can change things on the fly, and it’s live within minutes. With our team being smaller, it has made our lives exponentially easier.”
– Adelaida Nobles, Customer Experience Manager at Abra
They ultimately deflected 40.7% of support tickets, which eased the burden on their human agents and gave them more time to focus on high-priority tasks.
Example 3: Online course platform handles a huge influx of customers during the pandemic
Kajabi is an all-in-one platform for creating and selling online courses and digital products. When the COVID-19 pandemic hit, there was an unexpected massive interest in online learning. They knew within months of the pandemic that they had to scale quickly and strategically—but were really careful about choosing a solution that wouldn’t decrease CSAT.
Smart self-help tools powered by AI have always been part of their long-term strategy. Now, they needed to accelerate their roadmap with a trusted partner. They chose Solve, which was able to deflect customers directly to existing knowledge base articles. Solve also began making workflows on its own to help Kajabi tailor support experiences for specific customer segments.
AI ultimately deflected 18 to 20% of almost 80,000 queries over a few months. They also used Triage to ensure that more complex, high-priority tickets were routed to the right agents, which has helped many other customers automatically route tickets without the need for manual work.
2. Routing customers to the right place
Nothing frustrates customers more than having to wait. Customer support teams know this, so they develop systems to ensure that customers get to the right person as quickly as possible.
For some, that means outdated routing tools that tend to get it wrong. The customer may get routed to an agent quickly, but they get ping-ponged back and forth to several people before landing in the right place. To avoid this, some companies have a team dedicated to manually routing tickets. This is usually more accurate, but it’s a whole lot slower.
So what’s more frustrating—a slow response or an incorrect one? In the end, you need both. The good news is that AI-powered routing tools can do both, eliminating the need for manual routing.
Example 1: Capital markets platform stops manually routing 9,000 emails a month
Q4 Inc. is a capital markets platform. Their Live agents handled 6,000 to 9,000 monthly support emails, phone calls, and other support channels.
Their manual ticket routing process had become a serious bottleneck. A mere eight queue managers had to interpret and categorize tickets for 60 support agents, but this process was broken and inefficient. Some agents were overwhelmed with tickets, while others had none.
Q4 Inc. used Triage to help repair the routing process. This AI-powered routing tool automatically categorizes and routes tickets based on content, context, and urgency with little to no training. Triage uses historical data combined with LLMs to predict and classify new incoming tickets.
It also predicts support ticket fields and uses Salesforce’s help desk to send tickets to the right agent. Within Salesforce, Q4 Inc. set up additional automation to assign cases based on agent availability. Now, cases are categorized based on skill level and bandwidth.
After standing up Triage, Q4 Inc. reduced FRT by 98% and improved CSAT by 20%.
Example 2: Shared EV provider prioritizes safety for its customers
Lime is the world’s leading shared electric vehicles (EVs) provider. They have support operations across over 120 cities in more than 30 countries, which means they handle many tickets in many languages.
Their ticket routing process started out manual, which was incredibly slow and occasionally resulted in a misrouted ticket. This risked not only customer satisfaction but also rider safety.
Triage creates the language and category tags and then works to determine the required service level based on the severity of the case. Lime has custom tags, and triggers set up so that when Triage predicts a specific tag, the case is routed to a specific queue based on those tags. Now, the most critical inquiries are handled first. For example, a question about what to do in an accident can now be prioritized over a customer who has a question about a charge from two weeks ago.
Triage now routes 98% of support tickets tagged automatically.
Example 3: A home decor company finds relief during the busy holiday season
Spoonflower is an on-demand company specializing in custom fabric, wallpaper, and home decor. It was struggling to manage the spike in tickets during the holidays. Its manual routing process became overwhelmed, slowing down its support team and frustrating customers.
When COVID-19 hit, ticket volumes grew even higher, and priorities were unclear. Each agent could decide whether a ticket was high, medium, or low priority, which sometimes led to a backlog of tickets across all priority levels.
They used Triage to eliminate all that subjectivity and manual labor. The AI-powered tool proactively predicts ticket characteristics, then prioritizes and routes them to the right team. Triage is able to prioritize tickets correctly because it understands the sentiment of each ticket. Patterns in text help it classify a customer’s overall emotions as positive, negative, or neutral and set the right priority before routing.
Triage predicts and tags support tickets with 90% accuracy and 98% coverage, nearly eliminating the manual effort required for ticket categorization.
3. Surfacing correct information while they’re live with customers
Your customers aren’t the only ones trying to find the right information to solve their problems—your agents are, too. There’s nothing worse than sitting on live chat with a customer who’s asked a tough question and unable to find the answer.
“Can you tell me how to link my online bank account to your investment tool? My bank isn’t on the list, but your system says I can do it manually by providing routing numbers. Where’s the field on the screen?”
Your agent knows the product team just pushed an update and that field moved—but where?
“I’d be delighted to help you with this today. Could you hold off for a moment while I find a solution?”
And the clock starts ticking. First, your agent filters through their inbox for the product launch documentation. Nothing about the routing field. Onto the knowledge base—but that hasn’t been updated yet. Next, it’s the internal wiki. There’s something about the field there, but it looks like it’s in a different place depending on location. Finally, your agent opens Slack to ping someone and waits for an answer.
Your agents are working through this process as fast as possible, but it takes time. Usually, it takes too much time from the customer’s point of view. AI can increase your CX team’s productivity by doing this instantly with a generative AI tool that works like their own personal assistant.
Example 1: An online learning platform replaces an outdated internal wiki
D2L is an online learning platform. Their support agents have a ton of information they need to access about the specifics of their courses for a wide range of clients, including teachers, students, administrators, and private companies.
They started with a knowledge management system that ended up slowing them down. It depended heavily on keyword searches, which often yielded a long list of irrelevant results. Agents spent more time searching for information than working on customer tickets.
They wanted to move faster but didn’t want to sacrifice their high CSAT scores. So, they used Assist, a generative AI tool that surfaced real-time, contextual suggestions while agents were live with customers. Each suggestion was based on different client types’ specific needs, significantly reducing the time agents spent searching for answers.
D2L’s support team reduced their ticket handle time by 13.7% and increased the average number of cases they closed per hour by 32%.
Example 2: E-commerce company makes it simple to find detailed technical information
Etekcity is an e-commerce platform for home improvement, consumer electronics, and outdoor equipment. Each of its products requires specific technical knowledge, and its support agents struggle to provide quick and accurate assistance for complex technical issues.
Agents mostly relied on knowledge shared around the water cooler. Otherwise, they’d have to manually sift through product manuals and a knowledge base while living with a customer. This meant responses were inconsistent, took longer, and weren’t always accurate.
They tried an internal chatbot, but it didn’t have the customization tools needed to support their wide range of products, so they turned to AI. Assist was able to instantly analyze those manuals and knowledge base articles to provide agents with with instant, accurate information in an interface that works like chat-GPT,
Junior agents could stop relying on senior agents for answers, which led to a 69.7% reduction in FRT and a 60% reduction in average time to resolution.
Example 3: Global workforce platform needed to train agents after an acquisition
Beeline is a platform that connects businesses with the global extended workforce. After a big acquisition, they were left with two support teams that operated in different countries. Historically, each team had been trained on different products and followed separate processes, which meant customers got different answers from different agents.
Beeline first attempted to build a homegrown chatbot to deflect tickets while training the new team, but they struggled to ramp up new agents quickly. So, they turned to AI to train their agents for them.
Assist was able to summarize and surface the right knowledge as articles, past cases, macros, and personal notes within Beeline agents’ help desk. For newer agents, having the right resources at their fingertips made answering tickets a whole lot easier.
Plus, agents can attach relevant help center articles to cases through Assist and find similar past tickets to include in their responses. By applying macros with placeholder text on the ticket, agents have a template with relevant information they can easily personalize for each customer.
FRT ultimately decreased by 24.3%, along with time to resolution by 52.2%. Assist also cut the time between case responses by 32%. Beeline was able to ramp up new agents much faster.
4. Automating reporting and insights
Sometimes, the simplest questions are the hardest to answer. Not because you don’t have the data—it’s usually the opposite. It’s hard because you have too much.
“How happy are our customers?”
“Did that new product land well in the market?”
“Why didn’t revenue grow last month?”
There are so many ways to answer these questions, and customer support touches them all. You could report on CSAT, FRT, your brand sentiment, and more. Today, many companies pick and choose what’s most important to report on because reporting is a lot of work. Not everyone has the resources or time to pull and analyze such massive data to answer these big-picture questions.
When we pick and choose the data we crunch, we don’t always get the whole picture. AI has the potential to eliminate the need for this choice. AI-powered analytics tools can analyze massive datasets without missing a beat, resulting in truly actionable insights to guide your business.
Example: A fitness company gets insights from massive data set without analysts
iFit is a global health and fitness subscription technology company with over 6 million members. Those subscribers collectively create massive customer interaction data from which the team wants to learn.
They processed this data manually by analyzing support tickets and customer feedback. It was time-consuming, and the team often missed trends, which meant they didn’t always react in time to customer needs.
Based on its analysis, Discover proactively recommends workflows that can be automated to maximize cost savings. After analyzing the intent of all tickets iFIT’s customers have logged in the past year, these recommendations are provided. Based on these recommendations, the support team opted to automate three workflows: “refund requests,” “membership cancellations,” and “equipment isn’t working.”
“Before Discover, we didn’t have a system in place to understand the top drivers on our member center. It was really difficult to understand why or if articles were deflecting chats and emails. Now, with Discover, I can easily understand which topics are most common and missing from our knowledge articles, filling in the gaps. It’s so easy to use, and having that data gives me confidence that I’m making the right decisions.”
– Dustin Auman, Operations Manager
Let your support team focus on what matters
Improving CX productivity isn’t just about answering more tickets; it’s about answering the right tickets faster and with better solutions. AI has come a long way in the past couple of years. It’s realistically taking on the workload that live agents have always found cumbersome and time-consuming. That means you not only get happier customers—you get a happier team, too.
Forethought’s AI-powered tools like Solve, Triage, Assist, and Discover are already helping real companies become more productive. You too can free your team from mundane tasks and empower them to focus on building customer relationships and solving the problems that matter most.
