This weekly update covers major developments in AI agents and agentic AI for customer service. While artificial intelligence is changing how companies help their customers, the journey is not as smooth as many businesses hoped.

The Big Problem with AI Customer Support

A large study by Qualtrics released on October 7 painted a mixed picture. The research looked at 20,000 consumers across 14 countries and found that overall customer experience quality is improving. Customer satisfaction reached 79%, and trust levels hit 76%. However, AI-powered customer service is struggling badly.

Nearly one in five customers who tried using AI for customer support said it provided no benefits at all. Even more concerning, half of all customers worry that AI will block them from reaching a real human when they need help. This shows a big gap between what companies want to achieve with AI and what customers actually experience.

Why AI Customer Service Is Failing

Isabelle Zdatny, a leader at Qualtrics, explained the main problem. Many companies view customer support as something that costs money, so they rush to add AI chatbots to cut costs. They think they can push customers away from expensive human agents and toward cheaper AI bots.

The trouble is these AI chatbots often don't work very well. Companies deployed their AI systems too quickly without proper testing. Some businesses fed their AI agents information from old databases without cleaning up the data first. One example: AI systems were pulling policies from the year 2001 because they couldn't tell which rules were current and which were outdated.

Zdatny said companies "got out a little bit over their skis," meaning they moved too fast. This led to really poor quality conversations between AI bots and customers. When customers have bad experiences with AI support, it costs companies money too. Half of bad experiences make customers spend less with those brands.

Major Companies Launch New AI Agent Tools

Despite the challenges, big technology companies announced exciting new AI agent products this week. On October 6, Oracle revealed new role-based AI agents built into their cloud applications. These AI agents are designed to help with marketing, sales, and customer service tasks. Chris Leone from Oracle said AI agents are transforming customer interactions from slow, manual processes into fast, proactive strategies.

Salesforce announced two important updates coming to its AI agents. First, the agents will gain the power of speech, allowing them to talk with customers over the phone and through websites. Second, they will use something called "hybrid reasoning" that lets them be creative when needed but also follow strict rules for important decisions like giving refunds. Adam Evans from Salesforce said companies need both flexibility and consistency. The voice features will be available starting October 21.

Zendesk joined the competition on October 8 with a new autonomous support agent. The company claims this AI agent can solve 80% of support issues without any human help. This is a bold promise in a crowded market where many companies are fighting for customers.

Success Stories and Best Practices

NewDay, a financial services company, shows how AI can work well when used correctly. They handle 2.5 million calls every year and added an AI-powered system to help their customer service team. The AI saves an average of 30 seconds per call, which equals the work of 10 full-time employees. It also helps the company manage an extra 600,000 calls.

The key to NewDay's success is that their AI supports human workers instead of replacing them. The AI gives instant help to customer service agents, speeds up training for new employees, and makes sure customers get faster service.

Experts agree this is the right approach. Research shows that by 2025, AI will be involved in 95% of customer interactions. Companies using AI well report 17% higher customer satisfaction, 30% lower costs, and even 4% more revenue growth. But the winning strategy is AI plus human agents working together. Machines handle repetitive tasks while people focus on empathy, nuance, and strategy.

Matt Trickett from Qualtrics said it simply: "Companies that design AI to support the relational, not just enhance the transactional, are offering the best customer experience". This means AI should help build relationships with customers, not just complete transactions quickly.

Looking Ahead

The future of AI in customer service depends on finding the right balance. Customers want fast answers, but they also want to feel understood. They expect personalized help, but they worry about privacy when AI systems use their data. Companies that listen to these concerns and use AI to support humans rather than replace them will likely succeed. Those that rush to cut costs with poorly designed AI chatbots risk losing customer trust and money.