Harnessing Generative AI for Customer Service: Bridging the Gap Between Enthusiasm and Expertise
Published on Jun 9, 2025
Earlier this year, I had the opportunity to participate in a panel discussion at ITEXPO in Fort Lauderdale, where we explored the role of Generative AI in Customer Service and Customer Experiences. As an engineer with decades of experience in machine learning—dating back to the early days of convolutional neural networks—and a business professional, I found the conference illuminating. ITEXPO, a key gathering for managed service providers who power the IT infrastructure of countless organizations, revealed trends and tensions that I’ve since seen reflected in statistics from Salesforce and others. My goal here is to share these observations, not as an all-knowing expert, but as a pragmatic guide, hoping to shed light on the path forward for our industry.
The Divide: Business Optimism vs. Technical Caution
At ITEXPO, Business leaders were quick to champion the transformative potential of generative AI—its ability to streamline customer service, enhance personalization, and drive efficiency. Their enthusiasm is understandable; the promise of AI is compelling. However, technical experts, those tasked with implementing these solutions, expressed a more cautious perspective. Having worked hands-on with AI systems, they emphasized the importance of selecting the right use cases and understanding the technology’s limitations.
This disconnect highlights a broader trend. According to Salesforce, workers using AI often feel less confident in their understanding of the technology compared to the executives making strategic decisions. This gap in knowledge can lead to misaligned expectations and risky implementations.
The Case for AI Pragmatism
As an AI pragmatist with a background in engineering and business, I see generative AI not as a magical solution but as a powerful tool—what I prefer to call “artificial cleverness” or, more accurately, machine learning. My perspective is informed by years of hands-on work across a diverse set of machine learning challenges, from image processing to broader applications. This dual lens allows me to appreciate both the potential and the pitfalls of generative AI.
One of the most striking takeaways from the panel was the audience’s hunger for clarity. Attendees, spanning business and technical roles, were eager for a starting point—a way to navigate the hype and misinformation surrounding generative AI. Too often, decision-makers lack the foundational knowledge needed to evaluate which applications are suitable, leading to a misalignment between customer expectations and the realities of implementation.
The Risks of Misunderstanding Generative AI
Generative AI’s capabilities are impressive, but they come with risks that are not always intuitive. For example, AI-powered voice transcription can generate entire sentences that were never spoken, introducing errors that could have serious consequences in customer service contexts. Similarly, while generative AI can outperform traditional OCR algorithms in speed and accuracy, it can also produce “hallucinations”—fabricated content that may go unnoticed without scrutiny.
These issues stem from the fundamental mechanics of how generative AI works. A basic understanding of these principles can empower decision-makers to identify where risks arise and which applications are better suited for AI. For instance, tasks requiring high accuracy and low tolerance for error may demand more human oversight, while creative or repetitive tasks may be ideal for AI augmentation.
Empowering Decision-Makers with Knowledge
You don’t need a Ph.D. in mathematics to make informed decisions about generative AI. A foundational understanding of its strengths and limitations can go a long way. Without this, businesses risk relying on vendors whose priorities may not align with their needs. How can you ask the right questions—about data security, confidentiality, or the risks of bias and hallucinations—if you don’t know where to begin?
For organizations without in-house data scientists, the challenge is even greater. Trusted advisors or educational resources can help bridge this gap, enabling leaders to evaluate vendor claims critically and align AI initiatives with business goals.
Moving Forward: Practical Steps for Adoption
The growing adoption of generative AI, as highlighted in industry reports, is both exciting and sobering. While confidence in AI is increasing, so are instances where misapplications lead to unexpected setbacks. To avoid these pitfalls, businesses should:
Invest in Basic AI Literacy: Equip decision-makers with a working knowledge of generative AI’s capabilities and risks.
Align Use Cases with Strengths: Focus on applications where AI excels, such as automating routine customer inquiries or generating personalized content, while maintaining oversight for high-stakes tasks.
Foster Collaboration: Encourage dialogue between business and technical teams to ensure strategies are both ambitious and grounded.
Evaluate Vendors Critically: Ask pointed questions about data security, bias mitigation, and error rates to ensure solutions meet your needs.
Conclusion: Balancing Ambition with Understanding
Generative AI holds tremendous potential to enhance customer service and experiences, but its success hinges on informed decision-making. By bridging the gap between business enthusiasm and technical expertise, organizations can harness AI’s power while minimizing risks. Let’s move beyond the marketing hype and embrace a pragmatic approach—one that empowers leaders to make choices that are as smart as the technology itself.