In customer service, building trust in AI is crucial for its effectiveness and long-term acceptance.
Two Ai implementations are already in play in the customer experience: Natural language processing (NLP) allows for conversational, human-like exchanges, while sentiment analysis ensures empathetic responses to customer emotions—a personalized, empathetic approach that also helps bridge the trust gap.
When deployed mindfully, AI can proactively anticipate customer needs, offering relevant solutions before issues arise, which reassures customers of the AI’s reliability and competence.
“When it comes to implementing AI in customer service workflows, start with use cases that have been proven safe to use today if implemented properly,” said Jonathan Rosenberg, chief technology officer and head of AI at Five9.
When using AI in customer facing use cases – such as voice and chat bots – organizations should build, train and test their models on real, business-specific customer data before deployment.
“Once these models are live, monitor containment rates, sample calls for quality, and implement regular tuning,” he said.
Unlocking AI’s Potential
Christopher Sladdin, director analyst, Gartner, said when designing Generative AI (GenAI) chatbot or interactive voice response (IVR) conversations, the focus should be on driving adoption and trust rather than containment.
“If the customer asks to speak to a human agent, or you determine that doing so would be beneficial in a sensitive situation or where a customer intent is emotionally charged, make it easy for them to do so,” he said.
He explained when organizations make it easy to transfer to a human, customers are far more likely to try the self-service option again in future.
Carmit DiAndrea, director, engineering, AI data management CX at NICE, added transparency is key.
“Clearly defined escalation paths and compelling success stories demonstrating reduced wait times or boosted resolution rates build confidence in AI's capabilities,” she said.
She added that emphasizing human oversight, with experts continuously monitoring and refining AI accuracy, mitigates concerns about unchecked automation.
From DiAndrea’s perspective, building trust in AI means only using AI that is purpose-built for customer experience (CX).
“This means only training AI on high-quality, industry-specific and business-specific data,” she explains.
Rosenberg said he agreed that for enterprises using AI for customer service, transparency and results hold equal importance.
“If customers can use AI tools effectively to achieve self-service, and they find these experiences positive, skepticism or distrust will evaporate,” he said.
He added transparency is a valid concern, and the most important part of it is ensuring consumers know they are talking to an AI, and not to a human.
“Transparency also matters in B2B relationships, where vendors of AI tools need to provide clarity to enterprises they sell to on what AI technologies are being used, how data is being handled, where it is being stored and how AI models are trained,” Rosenberg said.
Good Data, Good Results
AI is only as good as the data that it is trained on, so generic AI trained on the open internet will not generate accurate results.
DiAndrea noted AI must also be built with the proper guardrails to ensure that the AI speaks the brand’s language and stays within those guardrails ensuring only appropriate responses.
“Continuous refinement through regular updates and customer feedback is also vital for maintaining accuracy,” she said.
In a 2023 Gartner survey of nearly 1,000 customers and employees, respondents were almost equally split between those who would trust AI or AI validated by a human to resolve customer issues, and those who were more trusting of a human or a human validated by AI for this activity.
Considered another way, 87% of customers still wanted a human in the loop.
“While we’ve observed an increasing number of customers turning to GenAI tools to try and self-resolve their customer service issues, often this is happening before they interact with a company-owned channel,” Sladdin said.
If they can’t self-resolve their issue in these channels, then they’re less likely to want to engage with a company-provided GenAI capability or regular self-service afterwards – instead, they likely just want to get to an agent who can finally resolve their issue for them.
Sladdin advised businesses to keep this broader view of the customer’s service journey in mind when striking the right balance between automation and human involvement.
“Often, we don’t know what steps the customer has undertaken previously, but we do know that they still prefer interacting with a human,” he said. “Our role in service is to meet their expectations and mitigate disloyalty by resolving their issue in a low-effort manner.”
Long Term Customer Loyalty
DiAndrea said AI has the power to transform customer service from transactional to truly personalized.
“Maintaining consistency across all channels, whether AI-powered or human-driven, ensures a seamless and positive journey that fosters long-term trust and loyalty,” she said.
From Rosenberg’s perspective, GenAI will enable AI-elevated customer experiences that are fully contextual interactions – whether digital or with a human – based on all available data.
This includes customer profile, related accounts, billing status, buying history, browsing history, past issues and social media conversations, along with the emotional context of the current interaction.
All this data creates a hyper-aware, hyper-personalized context that enables the most appropriate response, including seamless hand-offs between human and digital.
“With proper human oversight to ensure accuracy, customers will feel well known and well taken care of, creating loyalty and trust,” he said.