The world has now run AI in customer service at scale for several years.

Some of the results are extraordinary. Some are cautionary tales told in courtrooms and reversed press releases.

For the last two years, the world’s largest brands have been running the biggest customer experience experiment in history. Banks, airlines, retailers, fintechs, and telcos across every continent have poured billions into AI-powered customer service; chatbots, voice agents, sentiment engines, automated resolution systems.

The results are now in. And they are far more interesting and more useful to CX leaders.

Some companies are seeing extraordinary results. Others are quietly reversing course, rehiring the humans they replaced, and learning expensive lessons in public. The gap between the winners and the losers is not about how much money was spent or how advanced the technology was. It comes down to a much simpler distinction: whether AI is being used to augment customer experience or to replace judgment.

The evidence is no longer theoretical. After billions of customer interactions, the patterns are becoming clear. Here are some of the companies shaping the conversation on AI in customer experience.

The Failures

The most cited cautionary tale in global CX right now is Klarna.

In early 2024, the Swedish fintech giant announced its OpenAI-powered assistant was doing the work of 700 customer service agents, handling two-thirds of all service chats. Resolution times fell. The system handled queries in 35 languages.

It looked like the clearest AI success story in financial services. However, the company later acknowledged it had focused too heavily on efficiency and cost, and began bringing human support back because service quality had suffered.

Klarna began rehiring human agents to handle exactly the cases AI had been given that it was never equipped for.

Example 2: Air Canada’s chatbot confidently invented a bereavement fare refund policy that did not exist. The customer relied on it, was later denied the refund, and took the airline to court.

Air Canada lost. The tribunal ruled that the company was legally responsible for what its chatbot told customers establishing a precedent now cited across the industry.

The Successes

Erica, Bank of America’s award-winning, AI-driven virtual financial assistant. It has surpassed 3 BILLION client interactions!

Rather than acting as a bottleneck as so many chatbots do, customers can either complete the task within Erica or Erica will put customers on the best path to reach their goal including passing the torch to human representatives.

Australian health insurer nib Group Limited deployed AI-driven digital assistants to support customer service. The initiative generated approximately AUD $22 million in savings, reduced customer service costs by 60%, and lowered calls routed to human agents by 15%.

Lessons from the World’s Biggest AI Customer Service Experiments

1. Don’t confuse efficiency with exceptional customer experience.
Klarna proved that AI can dramatically reduce costs and speed up service. However, when efficiency became the primary goal, service quality suffered, forcing the company to bring human agents back. Saving money should never come at the expense of customer trust.

2. Your AI speaks on behalf of your brand.
Air Canada’s chatbot invented a refund policy that didn’t exist, and the airline was held legally responsible. The lesson is clear: every response your AI gives carries the same weight as if it came from one of your employees.

3. AI works best when it knows when to hand over to a human.
Bank of America’s Erica is successful because it doesn’t try to do everything. It resolves what it can and seamlessly directs customers to the right human support when needed. Great AI knows its limits.

4. The biggest wins come from augmentation, not replacement.
Australia’s nib Group achieved significant savings and reduced call volumes by using AI to support customer service not eliminate it. The most successful organizations use AI to empower both customers and employees, creating better experiences while improving operational efficiency.

Conclusion: The Future of Customer Experience Is Human-Led, AI-Enabled

The first wave of AI in customer service has taught us an invaluable lesson: AI is not the future of customer experience on its own. The future is the thoughtful partnership between human intelligence and artificial intelligence.

Organizations that pursued AI solely to reduce headcount are now revisiting their strategies. Those that invested in using AI to empower employees, remove friction, and create better customer journeys are seeing measurable improvements in efficiency, satisfaction, and loyalty.

The question for every business is no longer “Should we adopt AI?” It is “How do we adopt AI in a way that strengthens the customer experience rather than weakens it?”