The GPT Effect: A New Era of Customer Service

The world of customer service is rapidly changing. In 2022, support teams found themselves balancing headcount and budget restrictions while coming up against increased customer expectations and demands. Recent artificial intelligence (AI) breakthroughs have exceeded expectations and opened up exciting new possibilities. 

We are at the cusp of the biggest transformation in customer service in decades, and the future will see customer service that is more powerful and affordable for businesses, leading to more satisfied customers and happier, more efficient support teams.

The Rise of ChatGPT

OpenAI’s ChatGPT had over one million users within one week of its launch—and the implications for customer service were immediately obvious. For years, the industry has lauded the potential for AI and machine learning to transform the way support agents work radically, but the rate at which it’s starting to happen has taken everyone by surprise. 

Fundamentally, most customer support teams today operate as they did in the early 2000s, and we haven’t seen true innovation in the support space in decades. The introduction of Large language models (LLMs), like the recently launched GPT-4, has brought a new wave of opportunity and change to the customer service industry. LLMs have been trained on vast amounts of text data to understand the patterns and structures of language and can generate new text that resembles human language. In customer support, this can look like automatic language translation, text summarisation, rephrasing, tone editing or answering questions in conversations between businesses and their customers.

AI has the potential to improve customer support agent productivity drastically and free up their time to spend on more complex issues, pointing toward a future for agents to develop new skills and create long-lasting careers. But with all this innovation, there are kinks to work out. Until recently, AI-powered bots were not confidently ready to be put in front of actual customers, let alone solve their queries. And despite their potential, the latest LLM models “hallucinate”, meaning they make things up. And they do this frequently and convincingly —as seen when Google’s Bard produced a factual error in its first demo.

However, the recent launch of GPT-4 in March has changed everything—this updated model has been designed to reduce hallucinations, and AI has finally crossed the threshold to behave predictably. This means very soon, we will see most customer interactions assisted by smarter and more conversational bots that will route conversations to the right team, speed up teammate response times, parse through help centre content to respond to customers, respond to customers in a more human-like tone and enable teams to understand the trends in their conversations—making it primed and ready for the customer service spotlight.

Integrating AI into Your Support Strategy

Automation and AI alleviate some of the most common pain points for customer service teams—handling high call volumes, answering repetitive queries and dealing with disgruntled customers. These factors contribute to declining support team morale, but 81% of support leaders believe technology will help improve the employee experience and reduce attrition. 

Support leaders are looking to not only relieve the pressure on their teams and make agents’ jobs more rewarding and interesting, but they are also looking for strategic ways to improve the customer experience, and that requires a blend of humans and automation. Here are a few benefits of an AI-first support strategy.

• Improve agent productivity

LLMs are fantastic at reformatting or reprocessing text that’s already written, so they are perfectly suited to rephrase, expand upon or summarise the text in a support conversation. AI-support tools can enable agents to send faster and better quality responses to customers and reduce the time they spend composing messages or handing over conversations. 

AI will play a critical role in reducing agents’ time on rote work, allowing them to focus on the more complex issues at hand and increase overall team efficiency.

• Reduce support volume and resolution time

With the latest AI models, it is feasible to look at AI accurately and immediately answer 50%+ of customer questions, significantly reducing the time that supports reps spend answering repetitive questions. 

At Intercom, we have seen the impact of these new AI models first-hand in the form of our new AI chatbot, Fin. It works “out of the box,” using answers based on existing help centre content to instantly reduce support volume and resolution times without needing weeks of training or taking up precious agent time. It can understand queries across multiple conversations, allowing customers to ask follow-up questions and get additional clarification.

Inevitably, chatbots won’t be able to answer all customer queries but can pass the more complex, nuanced questions along to human support agents seamlessly. 

• Increase Customer Happiness

With retention top of mind for many forward-thinking businesses, providing world-class customer support has never been more important. By leveraging chatbots and automation to deliver instant answers to simple questions and by empowering support reps to handle the more complex queries with an empathetic human touch, brands can deliver personal service that ticks all the right boxes and keeps your customers engaged and happy. 

In the future, as AI and automation augment the human support role, new support team career paths and opportunities will quickly emerge. As AI becomes more adept at responding to routine customer queries, human agents will be freed to focus on complex and nuanced queries. Formulaic support work will be automated, opening up a future where agents answer each question for the first and last time.  

AI presents a major opportunity, and the business benefits are already evident. Support leaders who aren’t paying attention to AI and automation are missing a huge opportunity and risk getting left behind. The technology is maturing rapidly, and if there’s ever been an opportunity to adopt it and impact how you serve your customers, it’s now. Don’t let your competitors beat you to it.