As is the case for most in-house legal teams, the pressure to do more with less has been ramping up for years for the lawyers at reviews site Trustpilot. So when Chief Trust Officer Anoop Joshi, who is also a software developer, heard about generative AI (GenAI), he immediately saw the potential to use it to answer some of the time-consuming customer questions that take his staff away from doing higher-value work.
Rather than look for ways to fully automate these interactions, he knew the company needed to keep a human element in the process. When it comes to legalities, the need for accuracy and context is too great to fully outsource to a technology, no matter how smart it is.
The team used a tool called Wordsmith to train its GenAI model on Trustpilot’s legal playbooks, which cover topics such as contracts and privacy, and integrated it into the company’s internal Slack system. That way, the AI will not publish its answer until it has been verified by the relevant legal team member to ensure the information is reliable and sanctioned to be shared.
“Rather than having to answer the same questions over and over, the AI tool becomes like a middle man. It’s very comforting to know that we will not respond with an answer unless it’s been verified by a relevant member of the legal team,” says Joshi. “So it’s not replacing our lawyers. It’s making our lawyers more efficient.”
Welcome to customer service in the AI era. For decades, IT leaders have been under pressure to use technology to automate wherever possible, to drive down the amount of costly human labor required to get a task done. Too often, the customer’s experience suffered, as anyone who has interacted with a telephonic chatbot that won’t connect you to a real person can attest.
Yet while GenAI is the most promising technology for automating jobs that used to require humans, forward-thinking companies are following Joshi’s lead. They are using GenAI not to push humans out of the loop, but to create workflows that meld AI and human skills.
In fact, the invention of new ways of working optimized for human-AI collaboration is one of the hallmarks of enterprise AI maturity, according to research from ServiceNow and Oxford Economics.
“The advantage of AI is not necessarily what it does on its own, but the cognitive offload that it offers for your people,” says Chris Chiancone, chief information officer for the City of Carrollton, Texas.
By using GenAI to respond to repetitive requests to reset passwords, reschedule meetings with town officials, or approve invoices, “our employees can focus on more complex problem-solving, creative projects, and tasks that require human judgment and empathy. This shift not only enhances productivity, but also improves job satisfaction,” he adds.
To that end, Chiancone is developing a new workflow that lets city hall deliver services directly to consumers far more seamlessly, in a far more conversational manner. If a resident owes a court fee or may face a higher water bill because of higher-than-usual water usage in the area, they will be contacted via phone by a GenAI-based application that will discuss the issue with them.
If the resident says they can’t currently afford the court payment, the technology may suggest a six- or 12-month payment plan. If it hears a keyword or phrase such as “talk to an agent,” they will be automatically transferred to an agent.
AI can even enhance customer service reps’ability to behave more humanely, says customer experience author and advisor Adrian Swinscoe. He notes that Humana, the healthcare giant, uses a conversational analytics tool called Cogito that prompts reps to respond with more empathy—for example, to speak slowly when talking to older customers. After agents began using Cogito, the customer satisfaction rate went up 28% over three to four months, the company reports.
Others use AI as a triage or screening tool, says customer experience consultant Andrew Charles Moorhouse. The software acts as an “intent-level orchestration layer” that routes customer interactions to the best place for resolution.
Chiancone, for example, hopes to develop an improved version of a 911 operator that would screen calls based on a complex set of rules. The AI, while presenting itself as empathetic, could very quickly determine if it could and should handle a request. “If it’s a simple query, the AI might be able to handle the whole call. But if it’s ‘my husband’s having a heart attack,’ a human would be automatically called in to take over,” he says.
Customer service leaders have recognized the promise of AI. According to a recent study by McKinsey, more than 80% of respondents are already spending on GenAI or expect to do so in the coming months. These investments include not only AI systems that interact with customers directly, but also tools to support human agents when those systems can’t completely resolve an issue.
“AI-based agent support systems are already becoming a key tool for companies seeking to offer extremely effective personal service to demanding customers,” the report notes.
That resonates with Michael Park, head of AI go-to-market at ServiceNow.
“If I can drive the self-service or deflection rate up, it means that I'm delivering what those individuals need at greater speed and with greater accuracy, so they’re going to be more satisfied. But you won't be able to fulfill all of those at 100 percent,” Park says.
“Through the application of AI case summarization, knowledge base creation, and the automatic generation of notes, you can push productivity to new heights.”
The benefits of improved customer experience may outweigh any additional cost. According to a separate McKinsey report on the banking industry, AI could provide up to $1 trillion of additional value in this sector alone when cost savings and added revenues are considered.
Fred Reichheld, inventor of the Net Promoter Score measure of customer satisfaction, has long argued that good customer service is good for business. He found that a 5% increase in customer retention produces a more than 25% increase in profit.
There will be plenty of opportunities for further advances that will benefit customers and customer service providers alike. Swinscoe urges companies to use AI imaginatively: “Our ability to navigate in the future will be determined not just by the technology and technological advancement, but by how creative, imaginative, empathetic, and understanding we are, and who we do it for.”
Find out how ServiceNow can help you put AI to work for customer service.
