Q&A with legal expert on proposed NYC customer service chatbot law
New York City may pass legislation imposing guardrails on automated customer service agents, with consequences for retailers there using the technology.
Chain Store Age recently spoke with Alexei Klestoff, senior legal director at law firm ZwillGen in San Francisco, who focuses on technology and Internet law, privacy and AI, about a law recently proposed by the New York City Department of Consumer and Worker Protection (DCWP) to help ensure fair consumer practices by companies that use customer service chatbots.
[READ MORE: Poor automated customer service leads some consumers to switch brands]
Klestoff previously served as director of consumer protection at Walmart Global E-commerce, where he advised on consumer protection and privacy risk across the business, and is a former litigator at Morrison & Foerster.
How would you describe the proposed New York City chatbot law?
Essentially, the law is aimed at the so-called “doom loop” of automated systems where the customer gets stuck in a in a circle of unhelpful responses during automated customer service. It considers it to be an unfair practice to have a customer service line where customers repeatedly route through irrelevant menus, or a customer service function that provides wrong or irrelevant or unresponsive information.
The way to avoid these results is to provide customers a simple way of reaching a human, such as pressing zero or saying ‘representative.’ The law also goes into the details of how a company would provide immediate assistance if a human customer service representative. There are a variety of different ways the law would allow a company to satisfy its customer service obligations.
The law also requires customer service documentation. Companies would need to keep customer service records for at least three years.
If you are doing business in New York City and you're running a customer service bot, what would you need to change if the law took effect?
You need to have an easy way for customers to reach a human. You should audit your customer service processes to make sure that there is a way that someone could exit the automated process and actually speak to a human.
Obviously, most companies won’t be able to have a human readily available 24 hours a day, seven days a week. You need to evaluate different contact methods, such as letting the customer know what hours human assistance is available or have a human agent reach back out to them at a later time.
How does the proposed New York City law compare to the recently passed California automated customer service law?
Most of the provisions of the California law don't go into effect until January so there haven’t been any immediate consequences. But it’s narrower than the New York City law in some ways and stricter in other ways. While the New York City law applies to virtually all businesses, the California' law applies to large private businesses, which are businesses with more than $500 million in annual gross revenue.
But the California law has stricter standards for response times. During business hours, you have to make a good faith effort to connect a customer with a human within 15 minutes, and if you can't do that, then you need to allow them to schedule a time when they can talk to someone within one business day.
There are some exceptions, such as for asynchronous channels like email and online forms, but fundamentally no matter what the communication channel someone must be able to get a human to respond.
For a national retailer, would it be easier to comply with all local automated customer service regulations at a systemwide level?
The tricky part here is going to be the staffing piece. Compliance may mean that you need more human customer service reps rather than fewer, which would eat into the cost savings of automated tools. I would imagine that it would be easier to deal with these two jurisdictions separately, unless there is a groundswell of these laws.
Do you think laws like these might dampen enthusiasm for using customer service chatbots?
I don’t think these laws would completely eliminate the desire to automate customer service. They will just require additional workforce planning. The fundamental consequence will be the cost savings of automated customer service tools are probably going to be less.
How will the proposed law be enforced?
Neither of these laws allow a private right of action, so companies are less likely to get hit by something like a consumer class action suit. Enforcement tends to come from government regulations. In California, the penalties are $5,000 for an initial violation and 10,000 for any additional violations. The New York City law has escalating penalties, but I think the highest level, for three or more violations is $3,500 per violation.
