AI in Business: Effectively Implementing Generative AI in Customer Service

When OpenAI released ChatGPT to the public at the tail-end of 2022, it marked a significant milestone in AI development. Arguably the first generative AI to be embraced as mainstream by both professionals and the general public, its release sparked a wave of excitement, interest, and fascination with how new technology might change the world as we know it. It also evoked a strong reaction: fear. How will the integrity of education be impacted when students can now use generative AI to answer test questions and write papers? What will the long-term implications of AI be? Lastly, this is the most pressing concern for professionals all over the United States, including customer service representatives (CSRs): will AI take my job??

As generative AI has sprung onto the scene, across sectors including business, education, manufacturing, transportation and logistics, and many more, this concern has become front and center. Generative AI is disrupting the way information spreads, and how humans interact with computers and is transforming how jobs across industries are done. According to the authors of the Harvard Business Review (HBR) article “Generative AI Will Enhance — Not Erase — Customer Service Jobs,” AI will transform, rather than replace human jobs. These authors have written books on AI and machine learning, held director and C-suite positions at large organizations, and are leading the charge on generative AI implementation at the companies they lead, making them credible experts. To summarize one of their key ideas, people should not worry about AI taking their job. They should worry about someone who understands AI taking their job. Generative AI is a tool that can be used by professionals on the customer service side of business.

Generative AI, when effectively implemented with strategic oversight and human collaboration, empowers organizations of all sizes (from a mom and pop coffee shop asking for accounting tips, to a Fortune 500 industry powerhouse) to revolutionize customer service. It does this by augmenting human capabilities, enhancing efficiency, fostering innovation, and maintaining trust and accountability across industries.

As Olivia Briggs , a recent Summer Intern at Translator, Inc. put it, “[My company is] only using generative AI as a tool to help in the process, and aid humans in facilitating conversations.”? Translator Inc. is an HR training company that uses AI to analyze potential sticky organizations involving employees that can come up in an organization. Customer service tasks will be broken down into those that can be automated, augmented, or remain human-driven, and as HBR put it, "The net effect on jobs will be to create a new set of human work tasks — many of them of higher value.” So how can businesses utilize this tool effectively in their customer service??

As laid out in Chief Information and AI Officer Javier Campos ' book, “Grow Your Business With AI,” this technology and its implementation into customer service and other business functions can be viewed through the lens of three phases-- human-led AI, human-AI collaboration, and AI-driven human augmentation. In the human-led model, AI is a resource for humans. It assists in decision-making by processing data and offering insights, and at this point, control over AI systems remains predominantly with humans. It can perform data processing and analysis at speeds far beyond human capability to generate reports and highlight key trends or potential issues. AI also helps with simple, rule-based tasks, like updating records regarding customer service calls, sending follow-up emails, or inputting call data into a customer relationship management (CRM) system, that are automated, allowing humans to focus more of their energy on decision-making and tasks that involve creative thinking.

In the human-AI collaboration model, AI “assumes an active role alongside humans, contributing its strengths and balancing human limitations” as it evolves. In this phase, AI is actively working alongside humans, providing recommendations or assisting in complex decision-making processes. For example, in a customer service environment, it can assist in managing conversations, offering suggested replies, and providing real-time data, which enables a smoother customer experience with fewer bottlenecks, and it can help in live sales negotiations by suggesting the next best actions based on customer data and sales strategies.

Finally, in the AI-driven human augmentation model, AI acts as a cognitive assistant, expanding human decision-making and abilities. One major way generative AI can be used to improve decision-making and allow humans to work on more strategic, bigger-picture tasks include the following that is a far cry from what is available in the first two stages is, as Campos summarizes it, contributing to “idea generation, solution exploration, and identification of untapped options, thus fostering creativity and innovation” (Campos 2023). At this point, AI is not necessarily in the trenches with a team of CSRs on the phone, but is in the boardroom, contributing to big-picture strategy and being an advisor who can make recommendations regarding the financial management of the team, people management, and the overall vision of the company’s customer service department and either affirm or challenge the ideas and plans of the business’ executives.

Before we can do a quality evaluation of what it looks like for an organization to effectively implement AI in its customer service, we need to establish what role do CSRs play in the management and oversight of this technology. As established by the authors of the piece from the HBR, it will be the job of CSRs to ensure that AI systems understand and solve customer problems in ways that align with customers' values and intent, requiring human oversight to gauge emotional and moral alignment. Humans are still the movers and shakers of business and are leading the AI to solve business problems and advocate for solutions, and AI is not driving the business and organizing human capital in an organization. In this way, humans are the clear drivers and “owners,” walking AI on a metaphorical leash in the direction they see fit. CSRs will still be needed and valued in companies as they are responsible for verifying that AI solutions accurately interpret customer needs, solve relevant issues, and respect customer values, continually monitoring for misalignment. For instance, imagine a customer contacts an AI-powered chat support (who would have been a human CSR in the past) asking, "I need help with my upcoming flight." The AI could interpret this as a technical issue with booking and provide automated steps to manage booking errors. However, the customer may actually need information on health and safety protocols or luggage restrictions. Human oversight ensures generative AI handles routine tasks effectively while correcting errors in interpreting customer inquiries, leading to a more efficient customer service department that keeps customers satisfied.

As put by Justin Fineberg , the CEO of the software development startup Cassidy, which enables AI automation that is personalized to any business, “Implementing AI means creating a strategic vision for its use, training staff to work with AI tools effectively, and continuously evaluating its impact on business processes” (Fineberg). This definition of what it looks like to implement AI in one’s business on a broad level makes sense. Using CSRs to manage and oversee generative AI’s use cases, making sure it aligns with strategic vision and customer needs, aligns with the insights from HBR on how employees will use it in customer service.?

How does the way generative AI is implemented in customer service look different for small versus large companies? Ted Ladd , the Professor of Entrepreneurship and Innovation at Hult International Business School and Instructor of Innovation at Harvard University, who has written extensively on the subject of AI in business and how it is transferring the way organizations operate, had plenty of insights on this topic in his Forbes article, “For AI, Does Company Size Matter?” In the past, large companies had a big leg up on smaller competitors who did not have the resources in terms of people and money to implement complicated AI systems, but today, “Companies no longer need large IT departments with deep domain expertise or armies of external consultants to design and implement AI. Increasingly, smaller companies are exploring off-the-shelf AI products and solutions to meet their business needs.” It is true that large organizations still leverage distinct advantages in terms of dedicated resources, like substantial research and development (R&D) budgets, and established data infrastructures to support an array of AI-powered tools for customer service, which support more customized AI applications tailored to specific business functions. At the same time, as Ladd’s research highlights, AI readiness now hinges less on organizational size and more on aligning AI implementation with company vision and fostering trust among employees. This is in large part due to how, “today, smaller companies are deploying AI faster, more transparently, and collaboratively, thus accelerating trust in AI” (Ladd 2023). Imagine a small online clothing store that starts using AI to help customers find the perfect outfit. They introduce a simple tool on their website that suggests clothes based on what a customer has browsed or liked. Instead of just giving recommendations, the tool also explains why it chose those items, like matching colors or similar styles. The store invites customers to share their thoughts about the suggestions, helping to improve the tool over time, and they also post updates explaining how they use AI to make shopping easier without sharing anyone’s personal information. By being open and involving their customers, the store builds trust while showing how AI can make shopping more enjoyable and helpful.

Regardless of business size, involving employees in AI adoption helps mitigate apprehensions and fear around new technology, contributing to the acceptance of new technology in an organization from those who work in the business. Having buy-in from the people who make things happen for your company in operations, sales, marketing, logistics, HR, IT, and every other department, is crucial for realizing the full potential of generative AI’s transformative impact on both customer service and broader business operations. This is relevant for both small and large businesses, indicating that AI implementation now focuses more on strategic integration rather than the company's size.

Jonathan Counts , Senior Director of Digital Experience Strategy & Delivery at GCI Communication Corp. , a 2000+ employee telecommunications company based in Anchorage, Alaska, is on the front lines as far as strategy is concerned when marrying AI with customer service at his company. Tasked with developing and implementing comprehensive digital strategies that align with organizational goals, shaping and executing digital vision, driving innovation, and ensuring the successful delivery of digital projects and initiatives, Counts was quick to clarify a key distinction when it comes to utilizing generative AI in customer service.

He frames it like this: “We often say ‘AI’ when talking about a lot of different use cases, but there are two major ones to be aware of. It is important to differentiate what it means to use generative AI in customer service externally and internally.” External use cases include AI-powered customer service chatbots, automated email responses, and analyzing customer preferences and past interactions to suggest products or services, to name a few common applications. Using AI internally, which is the only way GCI is currently using generative AI, can look like generating summaries of meetings and action points, allowing employees to focus on the discussion rather than note-taking, and assisting customers with common IT issues (like resetting passwords or troubleshooting software). It can also listen to sales calls and give the sales or customer service representative live feedback on how to handle client inquiries and questions based on the tone of the customer. The generative AI can look back at data from past calls. The AI tells the rep what it has seen work well with customers when clients use an inquisitive, angry, frustrated, or excited tone.?

Generative AI is clearly a powerful tool for customer service, which is why human oversight with generative AI is paramount to protect an organization's brand and reputation– earlier this year, Air Canada, an international airline, was held liable for its chatbot giving a passenger bad advice. In that case, the company’s chatbot promised a discount to a customer that the company did not intend to be available. When the customer applied for the promised discount after their flight, Air Canada said their chatbot had been wrong, and wouldn’t offer the discount. In response, the customer sued, and the airline argued that their chatbot was a "separate legal entity that is responsible for its own actions,” but not according to the courts. “The British Columbia Civil Resolution Tribunal rejected that argument, ruling that Air Canada had to pay Moffatt $812.02 (￡642.64) in damages and tribunal fees.” Tribunal member Christopher Rivers read a statement explaining the decision: "It should be obvious to Air Canada that it is responsible for all the information on its website, and it makes no difference whether the information comes from a static page or a chatbot" (Yagoda 2024). This recent story is a valuable reminder for organizations and business leaders who are responsible for customer service strategy and implementing AI in this crucial area of business that intentional human oversight is paramount. If the chatbot had been overseen more directly by employees, this blunder, which cost Air Canada to the tune of $812.02 as well as credibility with some past and future customers.

The chatbot gave false information in this case because generative AI can “hallucinate,” which is when the generative AI produces information that sounds plausible or authoritative but is incorrect, misleading, or entirely fabricated. With the power of this technology also comes the responsibility to steward it well and take serious risks of false information, potential plagiarism, and questions of data security that can arise from implementing generative AI in customer service without proper oversight and controls in place.

As put by Matt Beattie , a Head of Data and Analytics (also at GCI), “While using generative AI in customer service offers significant benefits, faster response times and tailored interactions just being surface level examples, it also introduces important ethical considerations.” Transparency is key—customers should know when they are interacting with an AI rather than a human. Misrepresenting AI as human could lead to feelings of deception or frustration if the generative AI fails to meet expectations, especially with millennials and Generation Z, who are more prone than older generations to vote with their money and take a strong stand with their spending to advocate for social change. Additionally, disclosure allows customers to adjust their expectations—for example, knowing they’re interacting with AI might make them more patient with certain limitations or prompt them to seek a human for more complex issues. Being upfront about the nature of the interaction demonstrates ethical responsibility and fosters confidence in the technology's use.?

Protecting sensitive customer data is critical as well. When generative AI is used in customer service, it often handles personal information such as names, addresses, financial details, or even health data. Protecting this data is of the utmost importance to prevent breaches that could lead to identity theft or misuse. Businesses must ensure that AI systems adhere to stringent data protection regulations, such as General Data Protection Regulation (GDPR) or California Consumer Privacy Act (CCPA), and employ robust encryption and secure storage methods. Additionally, companies should limit data collection to what is strictly necessary for service and clarify how customer data is used and stored. Regular audits of AI systems can help identify vulnerabilities, and providing customers with control over their data—such as the ability to delete or review their information—can further enhance trust.

One last ethical implication to consider is in addressing potential biases that could lead to unfair treatment or exclusion. AI systems trained on historical data may inadvertently perpetuate or even amplify biases present in the training material. For instance, in customer service, an AI might prioritize certain customer demographics for faster responses or provide less favorable solutions to specific groups due to biases in the data. To address this, companies can use diverse and representative datasets during training to minimize unintentional discrimination. Ongoing monitoring and testing of AI outputs can help identify and rectify biased behaviors. For instance, ensuring that the AI offers the same level of assistance regardless of a customer’s language, accent, or location can help avoid exclusion.

By prioritizing these ethical principles, businesses can harness AI’s potential while fostering trust and maintaining accountability in their customer relationships.

Generative AI offers immense potential to revolutionize customer service, but its successful implementation hinges on a delicate balance of human oversight and strategic integration. By augmenting human capabilities, AI can drive efficiency, foster innovation, and maintain trust—critical factors in any business' success. From small businesses leveraging off-the-shelf tools to large organizations deploying advanced, customized systems, the key lies in aligning AI adoption with organizational goals while addressing ethical concerns and potential pitfalls, such as AI “hallucinations.” As seen in cases like Air Canada, human oversight remains indispensable to ensure AI aligns with customer values and company accountability. To recap, the key to successfully integrating AI into customer service can be summarized as follows: by embracing AI as a collaborator rather than a replacement, businesses can navigate the evolving landscape of technology and create a future where customer service is not only more efficient but also more empathetic and innovative.?

Works Cited

Balamurugan, S., editor. Impact of Artificial Intelligence on Organizational Transformation.?????

Wiley, 2022, https://doi.org/10.1002/9781119710301.

Campos Zabala, Francisco Javier. Grow Your Business with AI?: A First Principles Approach for Scaling Artificial Intelligence in the Enterprise. First edition, 2023., Apress,?

2023, https://doi.org/10.1007/978-1-4842-9669-1.

Beattie, Matt. Personal interview. 7 Nov. 2024.

Briggs, Olivia. Personal interview. 29 Oct. 2024.

Chen, Ying, and Catherine Prentice. “Integrating Artificial Intelligence and Customer??

Experience.” Sage Journals, Australian and New Zealand Marketing Academy, 16 May??

2024, journals.sagepub.com/doi/10.1177/14413582241252904.

Counts, Jonathan. Personal interview. 31 Oct. 2024.

Daugherty, Paul R., et al. “Generative AI Will Enhance - Not Erase - Customer Service Jobs.”?

Harvard Business Review, Harvard Business Publishing, 3 Apr. 2023, hbr.org/2023/03/generative-ai-will-enhance-not-erase-customer-service-jobs.

Fineberg, Justin. “The Quick-Start Guide to Implementing AI in Your Business.” Microsoft 365,?

3 June 2024, create.microsoft.com/en-us/learn/articles/quick-start-guide-to-implenting-ai.?

Ladd, Ted. “For AI, Does Company Size Matter?” Forbes, 5 Oct. 2023,?

www.forbes.com/sites/tedladd/2023/08/25/for-ai-does-company-size-matter/.

Yagoda, Maria. “Airline held liable for its chatbot giving passenger bad advice– what this means?

for travellers.” BBC, 23 Feb. 2024, https://www.bbc.com/travel/article/20240222-air-canada-chatbot-misinformation-what-travellers-should-know

“AI in Customer Service: Everything You Need to Know.” Salesforce, Salesforce, Inc.,??

www.salesforce.com/service/ai/customer-service-ai/#1715885026803_pxt. Accessed 14 Oct. 2024.