The Human Side of AI: Building Better Customer Relationships

For all the talk about AI making customer service faster and more personal, there is still a harder question to answer: Does the customer actually feel better served? An instant answer or perfectly timed recommendation may be useful, but usefulness is not the same as feeling understood. When a customer has a complicated problem, is frustrated or simply wants to talk to someone, efficiency can quickly become the problem.

That tension is becoming more important as companies bring AI into more customer interactions. Customers are increasingly comfortable with AI when it saves time or effort, but many still expect a human option when the stakes are higher. The challenge then is figuring out how technology can remove friction without taking away the judgment, reassurance and connection that make good service feel like good service.

In this article, members of the Senior Executive AI Think Tank—a curated community of leaders specializing in machine learning, generative AI and enterprise AI applications—explore that challenge from different angles. They discuss personalization and customer memory, where human involvement matters most, how AI can give employees more room for empathy, and how customer signals can shape better products. They also examine what it really means for AI to remember—and how leaders can tell whether an interaction leaves customers feeling understood, rather than simply processed.

“One thing I’ve learned is that personalization and feeling understood are not the same thing.”

– Sai Krishna Reddy Mudhiganti, Staff AI/ML Software Engineer at Samsung Semiconductor

SHARE IT

Make Personalization Feel Like Understanding

Sai Krishna Reddy Mudhiganti, Staff AI/ML Software Engineer at Samsung Semiconductor, works at the intersection of AI strategy, product innovation, governance and enterprise adoption. His perspective starts with a subtle but important distinction.

“One thing I’ve learned is that personalization and feeling understood are not the same thing,” Mudhiganti says. “AI can recommend the right product and still make the interaction feel cold.”

The solution, he says, is to make human oversight part of the architecture rather than an emergency fallback. 

“My recommendation is to design ‘human in the loop’ as part of the experience,” he says.

That means AI should carry forward the details customers should not have to repeat: what they asked previously, what went wrong, what was promised and whether the current issue warrants human attention.

For leaders, the lesson is to measure whether AI makes customers feel known—not simply whether it makes the interaction faster.

Design AI Around How Customers Feel

Gabriella Goddard, AI Leadership Director, AI Catalyst Leader™ and CEO of Brainsparker Ltd., brings more than 20 years of experience coaching executives and innovators across technology, finance, pharmaceuticals, aerospace and other industries. Her work focuses on combining AI with distinctly human capabilities such as creativity, curiosity, courage and connection.

“One of the biggest lessons from developing our AI Creativity Coach is that people don’t just respond to answers,” Goddard says. “They respond to how those answers make them feel.”

That changes the design brief. AI should not merely complete a task; it should reduce anxiety and build confidence. Goddard says Brainsparker invested heavily in defining its AI’s “identity, personality, emotional tone, conversation behaviors, communication style and brand beliefs,” alongside its knowledge and guardrails.

Her advice to leaders is simple: “Treat AI as an extension of your brand experience, not just another technology project. Every instruction you give it shapes how customers feel about your business, influencing trust, loyalty and whether they choose to come back.”

Close the Loop Between Customer Signals and Product Evolution

Tipu Swaran, Founder and Managing Director of Paalam Labs, is a technology strategy and digital transformation executive with two decades of experience leading enterprise change across Fortune 500 companies. He argues that the relationship does not depend on human interaction at every step.

“The fear assumes customer experience is defined by human interaction,” Swaran says. “I’d argue it comes from customers feeling understood and seeing the product evolve around them.”

AI becomes distancing when companies optimize in isolation. It becomes relationship-building when it connects what customers signal to what the business changes.

“AI can anticipate—sensing intent and unspoken needs rather than waiting for annual surveys,” Swaran says.

He points to Uber anticipating an unarticulated need and Spotify’s Discover Weekly learning from actual listening behavior.

“Neither distanced customers through technology; they built intimacy at scale,” he says.

The broader lesson is to treat customer data as feedback, not merely targeting fuel. AI should help the company become better at responding to customers over time.

“If AI can remember the basics about you without compromising safety and security, the transactional nature will quickly diminish.”

– Anisha Manvatkar, Sr. Technical Program Leader at NVIDIA

SHARE IT

Make Customers Repeat Less

Anisha Manvatkar, Sr. Technical Program Leader at NVIDIA, has more than nine years of experience spanning digital transformation, AI, program management and business processes, with leadership experience at NVIDIA, SAP and Capgemini. She sees customer memory as one of AI’s clearest opportunities to eliminate transactional friction.

“If AI can remember the basics about you without compromising safety and security, the transactional nature will quickly diminish,” Manvatkar says.

Once identity has been validated, she says, customers should not have to reconstruct the entire history of a case. 

“AI should be able to surface or remember relevant context on its own without the human having to reenter all the historical events of that case.”

That is increasingly important as customers move among channels and expect continuity. Deloitte’s research on connected consumers finds that personalization can improve perceptions of generative AI, while privacy and security concerns remain important barriers to trust.

Leaders therefore need to treat memory as both a CX capability and a governance responsibility. Remember what helps the customer, explain how that information is used and protect the customer’s control over it.

Protect the Moments That Define the Brand

Daria Rudnik, Team Architect and Executive Leadership Coach at Daria Rudnik Coaching & Consulting, is a former Chief People Officer and Deloitte professional with more than 15 years of international executive experience. She focuses on leadership, teams and organizational effectiveness in an AI-driven world.

“Customers reach out because they have a question or a problem to solve,” Rudnik says. “For many everyday interactions, AI may be the best solution.”

But smooth interactions are not necessarily the moments customers remember. 

“Brands aren’t remembered when everything goes smoothly,” she says. “They’re remembered when something goes wrong.”

That is why, she says, leaders should identify which customer touchpoints are too important to automate and where the brand voice needs to be visible. 

Her conclusion is especially useful for executives deciding where to deploy agents: “The choice isn’t about AI versus people—it’s about protecting the interactions that define your brand and build long-term trust.”

Use AI as an Invisible Assistant

Uttam Kumar, Engineering Manager at American Eagle Outfitters, brings deep retail technology experience across point-of-sale, order management, inventory and customer relationship management. His perspective centers on a familiar retail problem: Technology can become the customer experience if leaders let it.

“Retailers often make the mistake of using artificial intelligence strictly to cut costs, but its true power lies in elevating the customer journey,” Kumar says. AI can handle routine questions while giving floor staff and support agents more time to build human relationships.

“Technology should act as an invisible assistant that empowers teams to deliver warm, empathetic service,” he says.

That distinction changes the ROI conversation. Instead of asking how many employees AI can replace, leaders can ask how much more attention employees can give customers because AI has removed repetitive work.

“Leaders must focus on augmenting human empathy rather than replacing it entirely,” he says. “AI should smooth out friction in the background so genuine human interactions can shine at every key touchpoint.”

Build Clear Escalation Paths and Human Ownership

Venkata Kondepati, Manager of Data Architecture and Engineering at Ascentt, has more than 24 years of experience in cloud engineering, data platforms and enterprise software. His work includes customer identity, cloud operations and data engineering, making him particularly attuned to the role of context in enterprise AI.

“AI should make customers feel better understood, not more processed,” Kondepati says. “The biggest lesson is that personalization is not the same as relationship-building.”

He recommends using AI to remember preferences, anticipate needs, reduce friction and give employees better context. But “if every interaction feels automated, optimized and impersonal, trust declines,” he says.

His prescription is operational as much as technical: “Leaders should use AI to support human connection, not replace it in moments that require empathy, judgment or accountability.”

He recommends designing the customer experience with “clear escalation paths, transparent AI use and human ownership of important decisions.” 

This, he notes, will make service faster while making customers feel more valued.

Measure Whether Customers Still Feel In Control

Dr. Aditya Vikram Kashyap, Vice President of Firmwide Innovation at Morgan Stanley, leads enterprise innovation across finance, with expertise spanning AI integration, governance, data strategy and responsible technology adoption. His focus is on making innovation scalable without sacrificing trust.

“The mistake is using AI to automate the relationship when its real value is automating the friction around that relationship,” Kashyap says.

That distinction should change the metrics leaders use. 

“Leaders should measure success beyond containment rates and cost-to-serve,” he says. Instead, they should ask: Are customers repeating themselves? Fighting automation? Losing agency?

Those questions matter because personalization can cross a line when customers no longer understand why they are being treated a certain way. 

“Personalization without judgment quickly becomes surveillance with better marketing,” he says.

The goal, Kashyap says, is not simply to know more about customers. It is “to become demonstrably better at knowing when and how to serve them.”

Ask What AI Actually Learns

Charles Yeomans, CEO and Founder of Atombeam, brings more than 25 years of executive and investment banking experience and a background as a former U.S. Navy intelligence officer. His view introduces a deeper technical question about AI memory.

“The industry’s answer to this problem is memory bolted onto a language model,” Yeomans says, referring to retrieval systems that fetch past context and place it into a prompt. “That helps, but the model itself never changes.”

For Yeomans, the distinction matters because retrieval can still make an interaction feel like the system is repeatedly consulting a file rather than developing an enduring understanding: “It is still guessing from general training.”

He believes the fix is architectural.

“When learning actually modifies the system’s internal structure, the customer’s history becomes part of how it reasons rather than reference material it consults,” he says. “That is the difference between an assistant who reads the file before each call and one who has worked the account for years.”

For leaders evaluating AI, he recommends asking one key question: “When this system learns something, what physically changes? If the answer is a database row, it is retrieval, and the relationship will always feel like it.

“The goal is not to make customers feel the intelligence of the system, but the care of the brand.”

– Goran Paun, Principal and Creative Director at ArtVersion

SHARE IT

Design for the Emotional Residue

Goran Paun, Principal and Creative Director at ArtVersion, leads brand, digital strategy, UX/UI and product design work for enterprises and growing organizations. His human-centered design perspective brings the conversation back to what customers carry away after an interaction.

“Personalization alone does not create a relationship,” Paun says. “AI can know a customer’s history, predict the next question and produce the right answer, yet still leave that person feeling managed rather than understood.”

ArtVersion calls the result the “emotional residue”: how an interaction leaves the customer feeling. 

“Did it reduce uncertainty? Did it give them confidence?” Paun asks. “Was their question answered at the level they expected?”

That gives leaders a useful design test. AI should remove the effort surrounding a task without removing the human element from the relationship.

“Ultimately, a customer’s willingness to return reveals whether the intended experience actually worked,” he says. “The goal is not to make customers feel the intelligence of the system, but the care of the brand.”

Putting Human-Centered AI Into Practice

• Design AI to make customers feel understood, not merely recognized. Use customer history and context to eliminate repetition while preserving human judgment when the situation calls for it.
• Treat AI as part of the brand experience. Define its personality, tone, boundaries and approach to uncertainty with the same care applied to other customer-facing brand elements.
• Close the loop between customer behavior and product evolution. Use AI to anticipate needs, identify patterns and turn customer signals into measurable improvements in products and services.
• Build customer memory with privacy and security by design. Give AI enough context to create continuity without forcing customers to surrender control over how their information is used.
• Protect high-stakes and high-emotion moments. Determine in advance which interactions require human involvement, especially when trust, accountability or brand reputation is at stake.
• Use AI to give employees more room for empathy. Automate repetitive work so customer-facing teams can spend more time listening, advising and resolving nuanced problems.
• Create clear escalation paths. A well-designed AI experience should make it easy to move from automation to a human when the customer needs judgment or reassurance.
• Measure agency as well as efficiency. Track whether customers feel they have control, understand what is happening and can get human help when necessary.
• Understand what your AI actually remembers. Executives should know whether a system is retrieving context, adapting its behavior or changing its underlying model—and set expectations accordingly.
• Measure the emotional residue. Look beyond completion rates to whether customers leave interactions with greater confidence, less uncertainty and a stronger reason to return.

More Human, Not Less

The most important lesson from the Senior Executive AI Think Tank is that customer relationships are not made more personal simply because technology knows more about a customer. AI creates deeper relationships when it uses that knowledge to remove friction, preserve context, anticipate needs and give people better choices—including the choice to speak with a human.

The next phase of AI-powered customer experience will therefore be less about replacing human connection and more about designing for it. The strongest brands will use AI behind the scenes to make every customer-facing moment feel more informed, more responsive and, ultimately, more human.