Better Together: The Working Relationship Between AI Agents and Human Expertise


Share on LinkedIn

Generative AI is rapidly evolving, with large language models (LLMs) communicating and tackling tasks in powerful ways. For the customer service industry, companies are using various levels of AI and technology throughout their operations – ranging from predictive analytics, workflow automation, quality assurance and more. Specifically, AI Agents are designed to be autonomous, meaning they can operate without human intervention to a certain degree. They can make decisions, learn from data, and interact with their environment in a way that mimics human intelligence.

AI is poised to establish a new standard for customer experiences, offering services that are more personalized, adaptive, and satisfying for everyone involved. However, the deployment has not been perfect as issues with hallucinations and misunderstanding from bots become headline news. This not only raises concerns for businesses, but prompts them to rethink their strategy in which AI is not one size fits all. Bringing quality assurance to AI Agents is a critical step in ensuring AI Agents continue to evolve and learn from human interactions.

The Power of Incorporating AI into CX

As AI agents evolve, the human’s role evolves with it. The human element—empathy, communication, and understanding—becomes even more critical than ever to recruiters and teams in the training process as LLMs become more consumer-facing. Bots are evolving from simple features into proactive, intelligent problem-solvers that combined with human expertise are the future of customer engagement.

Across all industries – and especially customer service – a key component of AI adoption is building a foundation of trust with customers. Artificial Intelligence is greatly enhancing consumer experiences in many ways, but there does lie a concern: the risk of eroding trust. Instances where AI hallucinates or misinterprets emotions, responding in a disconcerting tone, can undermine consumer confidence. Current research indicates that AI still struggles to discern highly nuanced human emotions, especially those conveyed subtly through tone or facial expressions. This underscores the significance of human training and QA, particularly in interactions with frustrated customers whose responses may be nuanced.

The Evolution of AI Agents and the Role Humans will Play

To dive deeper into the future of customer service agents, Zendesk recently surveyed CX decision makers and human agents about the future of CX and the evolving role of agents.

Eighty percent of the respondents affirm that AI stands as the pivotal technology driving the future of customer service. Companies and consumers see generative AI not as a depersonalizing force, but as a tool to make interactions more humanized and personable when approached correctly.

Below are takeaways from the survey about the evolution of AI Agents.

  • The responses showed CX leaders believe that in the short term, human agents will act as a troubleshooter, and AI agents as a “traffic cop” — so teams will focus on upskilling their human agents, while primarily using AI for ticket routing and deflection.
  • A few years down the road, human agents will transition into deeper product and data expertise while relying on AI agents to drive a greater amount of routine customer engagement. The survey showed sixty two percent of CX leaders and human agents say that in three years, organizations will prioritize candidates with an aptitude for AI and tech.
  • Within five years, AI will become more than a tool—it will drive the majority of customer interactions, preemptively addressing customer needs and streamlining processes automatically as issues arise.

The survey found that overall, CX leaders feel human agents are irreplaceable and will always have a role in CX, backing up the notion that as the AI agents evolve, the human’s role evolves with it.

The Power of Quality Assurance

Unlike AI, humans are finely attuned to emotional subtleties honed through lifelong training in emotional understanding. While AI is making strides, human expertise remains invaluable in navigating such delicate realms. Zendesk’s training process of their AI Agents focuses on teaching the AI how to identify nuanced emotions and understand how to appropriately respond.

QA for AI Agents evaluates 100 percent of AI Agent interactions and uses AI to spot interactions that require human intervention, including churn risks, incorrect workflows, and knowledge center updates. When this process is repeated on a large scale, the AI system can find patterns and understand what responses tend to elicit a frustrated reaction from customers. This way AI Agents can eliminate using certain words, sentences, tones etc. if they see a pattern of frustration in customers’ reactions towards similar responses.

At Zendesk, every time a human agent intervenes into the conversation, the AI Agent is watching and learning. This human intervention allows AI Agents to progress faster and prevent hallucinations because they learn how to detect errors and respond appropriately to complex situations. To create a satisfactory customer experience, the AI must understand the nuances of human reactions and that people will respond differently from one another to their responses. This is why it is so important to put human experience and emotions first when building out this new era of customer experience (CX).

“We have a complex product that requires continuous agent training and Zendesk QA is essential for us to identify and fill knowledge gaps in the team,” said María de la Plaza, Head of Community Operations, SoundCloud.

The Transformative Opportunity for the Future of CX

AI will transform customer service to benefit every role in CX including customers, agents, admins, and CX business leaders. AI – and especially, AI agents – present an opportunity to address these dimensions of cost, volume, and quality and ultimately, transform the CX industry. This revolution will set a new standard for customer experiences, service that’s more personalized, adaptive and, certainly, more satisfactory for everyone involved.

Customer Experience leaders and agents are both preparing for what happens to human agents and the role of human interaction in customer service as AI’s applications to customer service expand into areas previously reserved for human work. Ultimately, companies have options as they decide which type of AI to use and where to use it across the agent, admin and customer experience. The future of customer service hinges not just on AI alone, but on the powerful and transformative collaboration between AI and human agents.

Lisa Kant
Currently Senior Vice President of Product Marketing at Zendesk, Lisa Kant is a storyteller and product marketer who has been working in the tech space for more than a decade. Prior to joining Zendesk, Lisa ran product marketing for ThoughtSpot, the leader in search-driven analytics and spent several years in product marketing for the Salesforce Platform. Lisa holds an MBA from the University of Chicago and a BA from Yale University.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here