The rising cost of human-led support is a frustrating reality in customer service. Is generative AI customer service the answer?
First, let’s start with the problem. It usually goes a little something like this.
The customer service team agrees on a service level agreement (SLA) that seems manageable at first. With enough time for customer service agents to quickly attend to every ticket, customer satisfaction (CSAT) is high.
But when support volumes skyrocket, a never-ending stack of support tickets accumulate. Rushing through one support ticket to get to the next, agents become overwhelmed. Costly agent turnover follows. Did you know it costs between $10,000–$20,000 to replace a call center agent?
On the other end, customers suffer long wait times and poor experiences. Conversations and feedback are lost in the mix. CSAT and net promoter score (NPS) take a serious dive. SLAs are missed, and senior leadership is wondering why.
The feedback loop that previously helped identify opportunities is broken, and your overall customer service quality declines.
Paving the way for AI-first customer service
Then, automation comes in to save the day — capable of handling more customer inquiries, faster. But leadership is still lost on how their investment in customer service is paying off, because there’s no clear measurement of whether the automation is “working” by delivering true value to the customer and the business.
Metrics like containment rate can measure if the bot is deflecting customers away from agents, but it doesn’t tell you if the customers are actually getting the help they need. So far, customer service chatbots don’t have the best reputation, with 77% of respondents in a Ipsos poll reporting that they found customer service chatbots frustrating.
To make things more complicated and confusing, the hype around generative AI is resulting in even more downward pressure from senior leadership. But it’s difficult to separate the buzz from reality and understand exactly how AI can transform customer service and provide that ROI leadership is looking for.
There’s a reason it’s so hard to understand and to identify exactly where AI fits in your customer service strategy. In order to effectively implement generative AI, we first need to flip the human-first customer service paradigm on its head.
This means shifting from human-first to AI-first. To be clear, moving away from a human-first customer service strategy does not equate to human-less. But it does mean you’ll have smaller numbers of humans focused on more valuable work. In fact, your people, process, and technology strategy will change entirely.
You’ll no longer need to scale your support function as your customer base grows. You’ll be able to better understand — and prove to the higher-ups — how automation is performing and the exact steps you need to take to keep improving it. Better yet, the success of your customer service feedback loop will strengthen your customer service organization and give you more sway in business decisions.
Want to find out how? Keep reading.
The AI-first strategy
We’ll start where any good customer service endeavor should — the strategy. The AI-first strategy should be held accountable to one overarching measurement: Automated Resolution. Automated Resolution is a fully automated interaction between a customer and a business that’s considered relevant, safe, and accurate. And, of course, it has to actually resolve the customers’ problem without the involvement of a human agent.
As you continue to increase the rate of Automated Resolution, standard customer service metrics will follow:
- Automatically resolving common FAQs saves agents for more complex conversations, reducing cost per ticket.
- Delivering the most relevant, accurate, and safe answer increases CSAT, customer lifetime value (LTV), retention, and decreases average handle time.
Success metrics like containment leave much to be desired, only measuring if a conversation requires a handoff to a human agent. It’s through the lens of Automated Resolution that you can fully understand the value of your AI.
At this point, you might be itching to find out how to get started. Again, shifting from human-first to AI-first doesn’t translate to human-less. In fact, machine learning experts suggest you approach generative AI implementation with a human perspective and onboard the model like you would a new employee.
That means thorough and attentive training — the better you train the AI, the better it will perform. Here are three things to teach your AI that Senior Engineering Manager in Applied Machine Learning, Gordon Gibson, goes over in the video below.
- AI literacy: Start by identifying all the documents across your organization — chat transcripts, knowledge base articles, conversations — to ensure they are an accurate representation of your business and customer service operations. The goal is to make that knowledge highly accessible, so it becomes the source of truth for your AI.
- AI performance: Audit the tools and systems your agents use to resolve support inquiries today, and see which actions can be taken with APIs. While Large Language Models (LLMs) predict the next word, they can’t actually take action on behalf of a customer, but they can connect them to a third party software that can help them take that action. This is why an API strategy is so important. It’s what will help you power more complex actions (apart from FAQs) and personalize the experience so that customers know you know who they are and that you care about taking actions on their behalf.
- AI expertise: You’ll want to teach your AI to access expertise across the organization. In order to translate that institutional knowledge from human expert to AI, you need to first identify the experts. Then, set up routing technology to connect them when needed, and set up the tooling that allows them to offer their expertise to train your AI. This will help allow the AI to not only automatically resolve customer inquiries and help them take action, it will actually be able to route truly unique inquiries to people in your organization that are best equipped to handle them.
The AI-first customer service team
The success of your AI-first customer service is two-fold: it should automatically resolve the most customer service inquiries, but it also needs to do this with the least amount of human effort. With AI on the frontlines, quickly ingesting, interpreting, and using data to not only interact with customers, but to generate automated questions and answers in the AI solution, you can restructure your customer service organization to fit the AI-first strategy.
Customer service teams will spend less time on the building side, and more time analyzing and optimizing automation to improve Automated Resolution. It’s giving employees the opportunity to actually evolve their careers, and go from roles like customer service agent to bot manager, or customer support advocate to conversational AI specialist.
The implementation and maturity of AI grows over time. And the more investment you put into it and restructuring your CX team, the better the results. Let’s dive into what this looks like:
- Content creation and assistance: At the very base level, AI writes the content or helps improve and create variations for content. In this case, an employee in a bot builder role would spend their time writing more high-value automation flows, ensuring content is consistent and on-brand. They will oversee generated content, train and optimize the chatbot based on insights, and build flows that enable the customer to take action and self-serve.
- Automatic resolution with generative replies: Here, AI can scrape support documentation and deliver answers to questions without manual training. At this level, someone in a bot manager role can identify more automation opportunities. AI is using the content you already have, enabling you to launch conversational AI in hours instead of weeks, putting employees’ focus on building action flows, optimization, and auditing transcripts to find more ways to improve the automated experience.
- Automatic resolution with generative actions: At its best, AI resolves inquiries without manual training, including using integrations to other tools to look up data or take action on behalf of customers. A customer service automation leader will oversee the project planning and implementation of new support programs, develop long-term automation roadmap, and work cross-functionally and deliver customer insights to business development teams. And while they’re focusing on high-level KPIs and elevating the role of the customer service organization within the company, AI is automating complex use cases and automatically leveraging customer data to make informed decisions.
Future state: Closing the customer feedback loop
What does customer service look like in this new AI-first world? Well for starters, you’re driving the cost effectiveness of your customer service organization by improving FCR and average handle time (AHT). You’re proving the soundness of your new AI-first strategy by supporting your customer service team from an investment perspective. You’re meeting your SLAs, Automated Resolution goals, and leveraging automation to protect your human capital.
How does generative AI make this possible? AI can quickly and accurately understand heterogeneous data compiled from different sources — like customer service transcripts or feedback surveys — at the speed and scale that’s needed, especially during periods of rapid growth. It can instantly generate reports on customer insights, ensuring no inquiry, conversation, complaint, or suggestion gets lost or disregarded. Clear visibility into this data invites C-suite leaders to reconnect with the customer, closing the feedback loop for good.
