Customer service was one of the first functions to generate tangible applications for generative AI.
Whether summarizing cases behind the scenes, assisting agents on the fly, or interacting directly with customers, the use cases for generative AI in customer service are numerous, and many companies are experimenting with or have already deployed GenAI in this way.
But which ones are achieving real results, how do they benefit and what lessons can we learn from their experience?
Verizon: AI-powered customer service agents increase sales
Telecommunications giant Verizon is gradually ramping up its use of generative AI in a customer service context.
In May 2024, the company announced a suite of new applications for customers and employees, including “Personal Research Assistant,” an AI conversational assistant capable of suggesting contextual responses to customer inquiries, and “Personal Shopper/Problem Solver,” which analyzes a customer’s profile to suggest who they might be and why they might be calling.
Since then, Verizon Consumer Group CEO Sampath Sowmyanarayan revealed that sales at Verizon increased by 40% thanks to new technology allowing customer service agents to sell to customers.
“We are doing real-time reskilling from customer service agents to sales agents,” he said. Reuters.
Verizon implemented GenAI technology in partnership with Google, and declared that it allows customer service representatives to respond comprehensively to 95% of queries.
The company trained a version of Google’s large Gemini language model on nearly 15,000 internal documents, which form the basis of responses used by customer service agents, saving them time and effort that would otherwise be spent tracking down the right details.
To guide its implementation of generative AI, Verizon convened an AI council, with Debika Bhattacharya, director of technology solutions. said Business Insider last year. It also published a set of AI principles to ensure responsible use of AI.
According to Bhattacharya, Verizon’s goal is to enable hyper-personalization through generative AI, “operating at scale but viewing each caller as a segment of one.” Everything the company does with GenAI, she said, is done with the goal of making customer interactions as seamless as possible.
Take away:
Verizon’s choice to retrain customer service agents as sales specialists and require them to devote more attention to selling is interesting, but reflects Verizon’s priorities for its business while illustrating how generative AI has freed up customer service agents to do other things.
This is not a use case that will apply to all businesses, many of which will likely prefer to double the level of service provided. Either way, it demonstrates that with automation taking over more time-consuming tasks, human know-how can be directed to areas where it is most valuable.
The most intriguing aspects of Verizon’s use of generative AI are the predictive elements, with AI analytics able to identify the nature and type of a customer request in advance. While this is arguably not foolproof, when it works, it surely helps customer support agents respond proactively and not just reactively – while saving time that would otherwise be spent establishing context.
ING: Conversational AI increases customer satisfaction and loyalty while preserving trust
“Find out the problem first, then scale it,” is the approach of Ayush Mittal, head of IT at Dutch multinational bank ING. describe the company’s approach to generative AI.
ING used an AI-powered generative chat solution to solve some of its key customer service issues, such as high volume (ING receives queries from 85,000 customers per week) and the need to offer support to customers outside of traditional working hours.
The company is no stranger to automated customer service, having experimented with the use of chatbots since 2017; However, these early chatbot experiences were rigid and only allowed customers to choose from a range of predefined options. By using large language models and natural language understanding (NLU), ING could create a smoother conversation: as Mittal says, “virtual agents have the potential to answer any question and provide a more natural conversation flow.”
Working with QuantumBlack, the AI arm of McKinsey, ING pilot a new GenAI assistant with 10% of customers in the Netherlands using the mobile app’s support chat feature.
In the first seven weeks of use, the chat agent served 20% more customers than the bank usually offers customer support, providing a model that ING could gradually expand across ten markets. Mittal said conversational AI also enables personalized and efficient interactions and has improved customer satisfaction and loyalty.
Generative AI is of course not without its flaws, such as the propensity to confidently provide incorrect information – which could be very dangerous for a financial institution and would be disastrous for customer trust. This is why ING has strict safeguards in place: ING’s risk stakeholders have been involved from the start of the process, real-time monitoring and auditing are in place, and an unreliable AI response will trigger human intervention.
The ING chatbot also cannot give advice on specific topics, such as mortgages and investment products. Like Bahadir Yilmaz, Chief Analytics Officer at ING, said Fintech financing“Introducing generative AI techniques to a business problem is only five percent of the work.
“Ninety-five percent of the work begins next. It’s important to build systems around AI tools and it takes a lot of effort.”
Take away:
Many executives have lamented the need to put strict compliance guardrails around the use of generative AI, fearing it could hinder innovation. However, ING has addressed these safeguards, making customer confidence in the implementation of GenAI the top priority.
It also provides peace of mind to company executives, who know they have already done their due diligence and do not have to worry about GenAI introducing elements of risk or creating a sudden reputation crisis.
As Andrea Del Miglio, senior partner at McKinsey, said: “Improving the customer experience represents a huge challenge for the banking industry, but there is also a huge risk. You can’t just release technology out of the box to do it.”
“Leaders need to ask themselves: What value are you adding to technology? What benefit can you add… to make it more useful to customers and better meet their needs?”
United Airlines: GenAI-enhanced flight stories increase customer satisfaction
United Airlines’ Every Flight Has a Story initiative uses SMS messaging and app notifications to provide customers with detailed reasons for flight delays, leading to an improved waiting customer experience.
The airline started this initiative several years ago, but recently introduced generative AI into the message creation process, enabling greater scale and freeing up staff to solve more difficult problems instead of editing templates.
With the introduction of GenAI in “Every Flight Has a Story,” customer satisfaction increased by 6%, as reported by Jason Birnbaum, CIO of United Airlines. said CIO.com.
United Airlines’ human “storytellers” still review messages to make sure they’re appropriate for the brand, but Birnbaum said United is becoming more comfortable with using generative AI.
“We worked hard to refine this model to take operational flows, notes from our operational teams, the crew and all these different data sources,” he said, “and let the AI take all of that data and create a more transparent, empathetic, decisive and clear narrative as possible.”
Flight delay notifications may include details about a plane arriving late due to runway construction, or an early warning about crowded security due to an NBA All-Star Game, as well as advice for travelers to arrive early and use the United Airlines app to streamline their experience.
Birnbaum said CIO.com that United has also developed LLMs for use in purchasing and manager-employee communications and is beta testing the use of GenAI to improve operations summaries for shift transfers.
The airline’s creation and training of generative AI models is facilitated by United Data Hub, a centralized data lake that consolidates United’s data sources and allows real-time access to data.
Take away:
The United Airlines use case here shows how generative AI, when implemented in the right way, can improve and expand an already effective initiative.
“Every Flight Has a Story” was not designed for generative AI, but it is widely expanded following its introduction. In February 2024, Inc. reported that United only provided the more detailed history of 15% of flights, but hoped to increase that figure to 50% with the introduction of GenAI.
United Airlines launched this initiative as a way to add humanity and nuance to delayed flight advisories, and the addition of GenAI has not reduced that humanity – thanks in part to ongoing monitoring by human “storytellers.”
The United case study also shows how companies’ work to get their data in order can pay off when it comes to implementing generative AI. The work to create United Data Hub provided a solid database for United’s introduction of external and internal LLMs, which are now used for flight status updates, purchasing, operations, and more.