The surge of generative artificial intelligence (AI) applications is spurring exciting innovations and consumer experiments, but it also worries many people who are concerned about data privacy or only being able to communicate with a company through a bot. These concerns are especially acute in industries where customer interactions and data privacy are critical, such as banking or healthcare.
Some level of anxiety typically accompanies breakthrough technologies, and it’s natural to worry about a technology that mimics human intelligence. As this new class of large language models has emerged, however, most companies have put model risk, accuracy of the model’s output, and ethical use of data at the heart of their risk frameworks. They aim to ensure responsible uses of new AI technology.
Less appreciated is the risk that companies will cede the customer experience to models and bots designed to extract value in the short term, not to foster long-term customer loyalty. Companies might increasingly pair traditional AI and machine learning models with generative AI to deliver messages and offers to customers in more human-like ways. If we are not careful, profit-seeking bots, algorithms, and predictive models could indeed lead to dystopian experiences.
Even in the world of AI, customer love should lead the way. Traditional metrics of customer sentiment, such as Net Promoter Score (NPS), may start to look different, but one premise will endure: Every interaction enhances or diminishes a customer’s perception the company involved.
Informing each decision with the goal of enriching customers’ lives will lay down a reliable route to an AI-enabled future that creates more value for customers, employees, and shareholders. In fact, early published results from researchers at Stanford University and Massachusetts Institute of Technology show favorable effects from the rollout of an AI-based conversational assistant tool to 5,200 customer support agents in several countries. Not only did the tool raise agent productivity by 14% on average, but the AI-assisted interactions had higher average NPS, and monthly agent attrition dropped by 9%.
Make It Personal
Orienting AI around customer love requires a fundamental rethink of objective functions. Most existing algorithms optimize around ROI for a particular moment rather than around an entire experience. AI-enabled customer engagement holds the promise of a company learning more from each interaction and finding more ways to create value for customers.
That’s a good sign because customers increasingly expect more personalized, relevant experiences and are open to sharing their data in return. Bain & Company’s latest survey of almost 30,000 banking customers in 11 countries found that the respondents who agreed that their bank personalizes the experience are more likely to reward it with a higher NPS. There’s a 123-point difference in NPS between respondents who strongly agree that their bank interacts based on knowing who they are and those who strongly disagree.
One way AI refines personalization is through digital assistants for customers, as shown by emerging efforts in banking and payments. Royal Bank of Canada uses an AI-enabled assistant called NOMI to personalize digital money management for customers. Features include timely tips pushed to clients, personalized budgets, and savings recommendations based on spending behavior and cash flow. In the year following its launch, the results were promising, with 50% more digital interactions for NOMI customers relative to the entire customer base, 93% more time spent on financial accounts, and 2% attrition of NOMI customers vs. 8% for their peers.
Generative AI digital assistants are also helping employees to strengthen their customer connections, reinforcing the places where a human touch may be a source of differentiation. Morgan Stanley Wealth Management, for instance, is rolling out an AI assistant to help its thousands of financial advisers better support their clients in a personalized way. The assistant combines search and content creation so that financial advisers can quickly find and tailor the right information for each client at any moment.
Large language models will enable a new era of personalization. Machine learning techniques already turn each customer’s pattern of digital interactions into a unique behavioral “fingerprint,” and recent AI advances will now enable these fingerprints to include speech and text interactions.
Help employees help their customers
Companies should start with a few no-regrets cases to get the organization comfortable with generative AI technology. These typically use AI to help employees who deliver aspects of the customer experience so that humans can vet the model’s output. Examples include suggestions to relationship managers for the next conversation with a customer based on recent engagement or providing specific actions for handling collections with customers facing financial hardship.
The next wave of cases would feature AI embedded in standard operating procedures for employees. Promising cases include predictive routing of a customer’s inquiry to the agent best equipped to handle a particular issue or real-time script recommendations for relationship managers. The technology will listen to a customer call in real time and help agents know whether their interactions are creating a promoter or a detractor Other employee-supporting features in the near future could include devising a personalized offer with images and text that evoke a customer’s favorite hobby or reminding a relationship manager to call on a customer during key life stages.
In a few industries such as retail, a fully AI-enabled front line is starting to support automated engagement directly with customers. Over time, this digital front line could deliver service with the same thoughtful empathy as traditional human front lines. Bots will engage with customers and learn to serve up relevant products and information just as the best employees have always done. The best uses of AI may even completely reimagine the overall experience.
In a time of high inflation and tight economics, some managers may be tempted to use generative AI technology only to cut costs and improve efficiency. That would be misguided. While generative AI has the potential to bend the cost curve in many industries, the greatest value will come to those companies that focus on enriching the lives of their customers.
This content was originally published here.