This article is sponsored by Interactions and was written, edited, and published in alignment with our Emerj sponsored content guidelines. Learn more about our thought leadership and content creation services on our Emerj Media Services page.
In the financial services sector, understanding customer sentiment in real time has become a crucial aspect of improving customer experience and satisfaction.
In a recent episode of the ‘AI in Business’ podcast, Ivan Edwards highlighted the challenges of assisting customers with self-service options and the need for innovative solutions. His team at Cadence Bank explored technologies that allow customer service representatives to virtually mirror the customer’s interface, whether on the desktop or mobile app, providing real-time visual guidance.
A multimodal approach ensures that customers receive comprehensive support, leading to higher customer satisfaction scores (CSAT) and a more personalized experience.
Emerj Senior Editor Matthew DeMello recently sat down with Kyle Hathorn, Director of Customer Experience and Strategy at First National Bank of Omaha, and Phil Gray, Chief Product Officer at Interactions, to talk about personalization in financial services. Their conversation centers on leveraging AI and data to enhance customer personalization and engagement while balancing efficiency, ethical considerations, and evolving consumer expectations.
Interactions provide customer engagement solutions through its IVA. FNBO provides a wide range of financial services, including retail banking, corporate banking, investment banking, wealth management, and more across eight Midwestern states.
Their unique experience brings practical depth to the conversation — Phil contributes expertise in leveraging AI and data analytics for nuanced customer interactions. In turn, Kyle provides insights into operational scalability and customer-centric strategies, balancing efficiency with personalization.
The following analysis examines three critical insights from their conversation:
- Adopt AI-driven, customer-centric interaction metric: Utilize AI tools for real-time analysis of customer interactions to prioritize engagement quality, tone, empathy, and customer effort over traditional metrics like call times, enhancing both satisfaction and representative performance.
- Differentiate interaction types for proactive analytics: Segment contact center calls by their nature (e.g., transactional vs. complex) to enable granular analysis and tailored approaches
. - Humanize data for personalized insights: Extract customer preferences and cluster them into personas using implicit data, such as language use or interaction patterns, to guide agents in value-driven conversations without restrictive assumptions.
Guest: Kyle Hathorn, Director of Customer Experience and Strategy, First National Bank of Omaha
Expertise: Customer Satisfaction, Customer Experience, Strategic Planning
Brief Recognition: Kyle Hathorn is the Director of Customer Experience and Strategy at First National Bank of Omaha. Prior to this, he spent over a decade at Ritchie Bros., culminating in his role as Senior Manager of Sales Strategies and Effectiveness. He holds a degree in Corporate Finance from the University of Nebraska-Lincoln.
Guest: Phil Gray, Chief Product Officer, Interactions
Expertise: Machine Learning, Natural Language Understanding, IVR
Brief Recognition: Phil Gray is the Chief Product Officer at Interactions, where he has been for nearly 17 years. He holds a degree in Business, Marketing, and Management from Michigan State University. Phil has previously worked with eLoyalty, Versay Solutions, and Nuance.
Adopt AI-Driven, Customer-Centric Interaction Metric
Kyle opens the conversation by talking about the need to improve contact center operations by shifting the focus away from traditional efficiency metrics. These include the length of the call (“call times”) and how long customers were on hold (“hold times”), to more meaningful engagement models that enhance customer experiences.
Given the challenges he’s seeing currently at FBNO, he especially emphasizes the importance of ensuring customers feel heard, which requires minimal effort but has a significant impact.
Leveraging AI tools to analyze calls or video recordings in real time can help customer service representatives improve their interactions by providing instant feedback on tone, empathy, and responsiveness. He advocates for developing new metrics that prioritize the quality of interactions, arguing that customers value positive and practical experiences more than the speed of resolution.
A more quality-based approach, Kyle believes, encourages a customer-first mindset, supported by AI-driven insights, to elevate overall satisfaction.
Similarly, Phill highlights the need to shift contact center metrics from internal efficiency, like containment, to customer-focused measures, such as effort and sentiment. He emphasizes simplifying customer interactions to align with their goals and reduce frustration.
Advances in generative AI now enable real-time analysis of factors like resolution and sentiment without extensive training, but growing concerns about AI privacy must be addressed. Phill sees an opportunity for AI to move beyond analytics to actionable insights and automation, transforming customer service.
Kyle argues that sentiment analysis is challenging due to regional and cultural differences in customer behavior. For example, sentiment models often misclassify frustration levels in certain populations, such as Latina customers, due to language nuances. Kyle highlights the need to refine these models to ensure sentiment scores accurately represent diverse customer demographics and experiences.
Differentiate Interaction Types for Proactive Analytics
Phil then highlights the need to distinguish between different types of interactions in contact centers, particularly in financial services. Fraud-related calls often involve complex, multi-step investigations that can’t be resolved immediately, and the outcomes may not always favor the customer. This complexity impacts customer experience expectations.
He criticizes the traditional approach in contact centers, where all calls are treated similarly without accounting for their nature.
Phil continues to note that this leads to flawed data, as “many companies don’t even really distinguish between the two. They just have some kind of blob that was a call with some scoring involved at the agent level.”
Differentiating structured, transactional calls (e.g., balance transfers) from complex issues (e.g., fraud) allows for more granular and proactive analytics, improving both insights and outcomes.
In response, Kyle emphasizes the importance of enabling customer self-service for efficiency and scalability but acknowledges limitations in applying this approach universally. He highlights that some services, such as corporate treasury, require distinct staffing and models compared to others, like credit card servicing, due to fewer self-service opportunities:
“So I think, to Phil’s point, we probably are the snake that’s biting our tail off is we are going after an old, archaic model that maybe doesn’t serve the customer to the best benefit in a lot of ways.
But we do think there are certain aspects where we would want enhancements, like our corporate treasury service center is completely staffed and modeled differently than our credit card servicing center, and it is just less opportunity to do that self-serve capability.
So we always try to look at: What is a customer journey? What are the true end goals? What are customers contacting us for, and how can we get them to that resolution the fastest for their benefit? And so we think of it differently based off of population.”—Kyle Hathorn, Director of Customer Experience and Strategy at First National Bank of Omaha
Humanize Data for Personalized Insights
Phil highlights that many organizations fail to use their data effectively for dynamic, personalized customer experiences. He emphasizes leveraging AI to analyze customer behavior, predict needs, and identify risks (e.g., switching providers).
Offloading routine tasks to bots can free up human agents for high-value work. He also notes the need for frequent reassessment due to rapid changes in consumer behavior and technology capabilities.
Phil then explains how data can be derived implicitly during customer interactions. For example, AI can deduce a customer’s language preference without explicitly asking, simply by observing the conversation. Similarly, customer interactions can provide insights, such as associating a phone number with a profile or inferring demographic details based on voice characteristics.
However, he notes the ethical considerations of profiling customers using implicit methods, such as age or gender inferred from voice or collecting biometric data. While these capabilities exist, they raise questions about consumer consent and privacy boundaries. This highlights the balance needed between utilizing data for personalization and respecting customer trust and preferences.
In turn, Kyle amends Phil’s thoughts on the importance of striking a balance in leveraging customer data. He suggests organizations should focus on clustering customers into personas to humanize the data and help agents engage in more meaningful conversations.
By using generalizations, agents can suggest relevant products — like wealth products for someone mentioning retirement — without overstepping or making overly specific assumptions about the customer.
The key, according to Kyle, is to add value without boxing customers into predefined categories, ensuring personalization remains helpful and not intrusive. Avoiding pre-categorization avoids alienating customers while still enhancing their experience.
