Giant outdoor gear and clothing retailer Patagonia is being sued for its use of third-party AI-powered software that listens to, records, and analyzes call center conversations with the company’s customers and business partners.

Plaintiffs in the class-action lawsuit claim Patagonia, through its use of Talkdesk’s cloud-based technology, is violating California privacy laws by allowing their calls with the company to be tapped, recorded, and used by the third party without their knowledge or consent. The 18-page suit, filed July 11 in California Superior Court, is one of a growing number of actions taken to test and define the legal boundaries of AI technology – particularly generative AI – in areas ranging from copyright and trademark to privacy and consent.

In this case, the plaintiffs – led by Michelle Gills from Willits, California – are claiming their privacy rights were violated. Gills said in the lawsuit that she called Patagonia’s customer service line in January and assumed that the call would be kept confidential.

“She did not know that Talkdesk was secretly listening to her conversations, nor did she consent to her conversations being intercepted, listened to, recorded, and used by Talkdesk,” the suit claims. “She would not have communicated with Patagonia if she had known this were the case.”

Capturing Calls and Texts

Talkdesk includes omnichannel capabilities, able to track conversations not only via phone calls but also through texts and other avenues. The technology also includes features like Experience Analytics, which transcribes all conversations between call center agents and customers, processes the customer’s speech and text patterns, and then analyzes the data to gauge their intent and sentiment, according to the lawsuit.

There also is Quality Management for monitoring and evaluating call transcripts, real-time calls, and screen recordings. It uses algorithms to analyze the customer’s tone and phrasing to measure their emotional state. Copilot – or Agent Assist – listens to conversations in real-time and automatically transcribes them via speech-to-text and natural-language processing capabilities, analyzing the communications, and guides the agents.

“For example, Copilot will automatically suggest relevant responses for agents in chats, emails, calls, and texts based on the content of customers’ communications,” the plaintiffs say. “Talkdesk saves all this information in the cloud and builds an ‘interaction history,’ which enables companies to keep track of customers’ prior conversations – even if those conversations occurred in a different medium. All of this data is stored on Talkdesk’s servers.”

The lawsuit wasn’t clear on exactly what Talkdesk technology was used by Patagonia.

Personal Data in Conversations

The storing of conversations in Talkdesk’s servers is a key sticking point. Customers contacting Patagonia expect their conversations with the company are confidential, especially given the amount of personal information that can be disclosed during them, such as addresses, credit card numbers, and other financial data.

Instead, all the information in the conversations is also sent to Talkdesk’s servers, with the data intelligence and AI company using the information for its own purposes, such as training its AI models and improving its services.

When calling into Patagonia’s customer service department, neither the retailer nor Talkdesk get people’s consent. Patagonia’s privacy notice is visible at the bottom of the website, but it never requests customers’ consent. It only encourages them to read the notice to see how their data is being used and how they can exercise their rights. It doesn’t mention the use of Talkdesk’s technology.

“And when customers create a Patagonia account, they are not presented with (and thus not required to consent to) Patagonia’s Privacy Notice,” the plaintiffs said.

Call Centers and Generative AI

Call centers are popular use cases for generative AI technology, with Talkdesk rival Zendesk arguing in a blog post earlier this year that the emerging technology makes it easier for companies to deliver better customer experiences.

“Integrating AI into call center software for small businesses, startups, and enterprises helps companies of any size or industry deliver more efficient customer service experiences,” the company wrote. “Using machine learning, natural language processing (NP), and automation technologies, AI’s potential is seemingly limitless.”

It’s being used for such tasks as summarizing calls to save agents times and boost their productivity, transcribe and analyze calls to help with agent training and development, and direct customers to digital channels to reduce the number of calls. Benefits range from increasing customer satisfaction and scaling operations to reducing costs and improving quality, according to Zendesk.

The momentum behind using AI in calls centers isn’t going to slow. Analysts with market research firm Fortune Business Insights expects the global call center AI market will jump from $1.95 billion this year to more than $10 billion by 2032, growing an average of 22.7% a year.

The Expectation of Privacy

However, as the market grows, organizations need to keep a focus on the expectation of privacy when customers call their service centers. The plaintiffs in the lawsuit said they know that their conversations with Patagonia are financially valuable to the company with the information – not only personal and financial data but also verbal and acoustic information – “is a form of currency. The value is well understood in the e-commerce industry.”

The information is used for training Talkdesk’s AI models and improving its services, which makes those services more valuable.

They noted a passage in the Harvard Law Review earlier this year noting that “personal information is an important currency in the new millennium. The monetary value of personal data is large and still growing, and corporate America is moving quickly to profit from the trend. Companies view this information as a corporate asset and have invested heavily in software that facilitates the collection of consumer information.”