AI

Bots, AI agents and IVRs: Why responsible AI is the next CX differentiator

The future of AI in retail voice channels will be defined by who earns and keeps customer trust.

Photo: Adobe Stock

December 9, 2025 by Mark Frumkin — Director of Customer Success, Modulate

In retail, every customer call is a moment of truth where agents and customer support reps can deeply understand and resolve a customer's pain points. Increasingly, that moment is mediated by artificial intelligence. AI shapes how brands listen, respond, and resolve. But as the technology accelerates, one factor is emerging as the real differentiator in customer experience: responsibility.

Consumers have clear expectations. Gartner found that 64% of customers would prefer that companies did not use AI for customer service, and more than half said they would consider switching brands if they discovered AI was being used without transparency. At the same time, PwC reported that while 53% of consumers trust AI to assist in customer service, that confidence drops sharply when personal data or high-stakes decisions are involved.

Trust now offers the competitive edge. Marketplaces and platforms need to be transparent with customers regarding how they are thoughtfully using AI with empathy at the core. And nowhere is that more vital than in voice channels, where tone, timing, and emotion carry as much weight as the message itself.

The bot revolution

A growing frontier in voice interactions is the rise of AI bots acting on behalf of consumers. These AI-driven bots are increasingly used to call retailers to request refunds, check loyalty balances, or dispute charges with the customer's consent. In legal terms, that triggers agency law, which means the AI bot is acting as an authorized agent of the consumer.

That raises the intriguing question: Can a company legally refuse to engage with an AI representing its customer? This is particularly relevant because many companies rely on synthetic voice detection as a key fraud indicator (think of deepfakes). However, with the advent of these agents, as well as customers using synthetic voices for accessibility or identity reasons, platforms now need to be much more thoughtful in determining the significance of hearing an automated voice. The key is understanding when AI is being used to support the speech of a human customer, to act on behalf of a legitimate customer, or to attempt fraud, and treat each case individually.

Forward-thinking retailers are adopting practical approaches. Some are deprioritizing bot calls in the queue since AI can wait more patiently than a human caller. Others are experimenting with AI-to-AI interactions, where automated systems handle basic bot inquiries while human agents focus on complex or high-value conversations. These approaches respect the customer's intent while maintaining operational control.

AI should augment, not replace human agents

There's no denying the efficiency AI brings. Some solutions will transcribe and summarize entire calls for supervisors to review and use as training opportunities, while other AI systems can provide automatic alerts about recurring customer complaints. But full automation shouldn't be the goal. In the retail setting, AI's real value lies in making human agents more effective and less stressed.

Agent attrition across the industry remains high, with call center turnover rates often hovering around 30% to 45% annually. The emotional toll of difficult interactions drives much of that turnover. AI can help by detecting heated calls and routing agents to easier tasks afterward - though this requires voice-native analysis, not just transcription of the words being spoken which most AI systems rely on today. Voice-native AIs can also more effectively flag moments when the agent might need additional support or coaching. One large retailer saw morale jump significantly after introducing micro-breaks recommended by data insights with just five extra minutes of rest per month per agent.

AI can also function as a real-time expert assistant, providing cues like, "This customer sounds frustrated — slow your pace," or alerting agents to region-specific compliance requirements. By pairing empathy with intelligence, AI helps agents deliver better, more human service.

Voice-native analysis: The next frontier

Many organizations still analyze call transcripts as if they were chat logs. That's a mistake. Relying only on post-call text analysis can delay actioning by days compared with seconds in the case of systems that process vocal tone in real time. That delay can mean the difference between resolving a problem gracefully or losing a loyal customer.

Voice-native AI can also enhance investigations after a call, flagging where a speaker's tone became heated or where a synthetic voice appeared. Those insights not only improve fraud detection and quality assurance but also streamline reporting, saving time and reducing fatigue for agents and supervisors alike.

Building a responsible AI roadmap

Many enterprises begin their AI journey with the wrong question: "How can we use AI?" The better question is, "What problem are we trying to solve?"

Start by identifying a specific, measurable challenge, whether that's reducing fraud, lowering churn, or improving first-call resolution, and work backward. Often, the simplest first step is organizing the company's knowledge base, ensuring agents and AI systems have access to accurate, current information.

That foundation enables more advanced applications down the road, from adaptive learning systems to proactive customer outreach. And as AI evolves, brands should look beyond large language models toward targeted, domain-specific AI built for high-stakes, real-time interactions.

The bottom line

The future of AI in retail voice channels will be defined by who earns and keeps customer trust.

Responsible deployments of AI solutions strengthen agent well-being, safeguard data, and ensure every customer feels respected. This is both good ethics and good business: building trust by solving the painful problems with the right solutions is the differentiator in the customer experience of the future.

About Mark Frumkin

---

Mark is Modulate’s Director of Customer Success, leading a team in partnering with customers in the gaming and retail delivery space, including Activision, Ubisoft, Rockstar Games and Fortune 400 companies. Mark oversees Modulate's customer success strategy, ensuring smooth onboarding and ongoing successful outcomes for customers, helping them to detect and prevent abuse and fraud in their voice channels and phone lines.

Connect with Mark: