Source: University of Göttingen/Mozafari
Should retailers continue the chatbot deception?
Consumers connecting with customer service online often initially fail to recognize whether they’re communicating with a virtual agent vs a live one. A new university study finds both benefits and drawbacks to full disclosure.
Among the negative fallouts from full disclosure, researchers from the University of Göttingen found consumers tend to react negatively when they learn that the person they are talking to is, in fact, a bot. This is particularly true if service issues are perceived as particularly important or critical. Researchers concluded, “Chatbot disclosure has a negative indirect effect on customer retention through mitigated trust for services with high criticality.”
The findings align with previous studies showing that consumers have a negative reaction when they learn that they are communicating with bots. Disclosing that the contact was a bot, however, leads to positive customer reactions in cases when the bot cannot resolve the customer’s issue.
“If their issue isn’t resolved, disclosing that they were talking with a chatbot makes it easier for the consumer to understand the root cause of the error,” said Nika Mozafari, a co-author from the University of Göttingen said in a statement. “A chatbot is more likely to be forgiven for making a mistake than a human.”
Given their inherent aversion to chatbots, many customers may opt to choose to connect with a live agent most times if given the option.
Resolving customer queries in a fully automated way promises significant cost savings for customer service departments and better accommodates 24/7 coverage. For customers, bots can reduce the time waiting for an agent to resolve less-complex issues.
Customer service bots are growing in popularity, although they appear to still be preferred by many for simpler tasks.
A Genesys survey from late 2019 found nearly three-quarters (73 percent) of respondents open to dealing with a voice/chatbot. Half (51 percent) say this is only when the issue is simple or transactional, such as checking account balances, resetting passwords or confirming order status. When issues are more complex, such as billing errors or missed flight connections, only 21 percent of the consumers surveyed felt comfortable dealing with a bot.
- Trust me, I’m a chatbot – Eurekalert
- Trust me, I’m a bot – repercussions of chatbot disclosure in different service frontline settings – Journal of Service Management
- Astute Publishes Original Research That Finds How Consumers Want to Speak with Brands is Changing – Astute/Business Wire
- Genesys Survey Shows U.S. Consumers Are Warming up to AI: Nearly 70 percent Describe Past Experiences With Customer Service Bots as Positive – Genesys/PRNewswire
Discussion Questions
DISCUSSION QUESTIONS: How transparent should retailers be about letting customers know they’re communicating with a live or virtual customer service agent? How would you rate the pros and cons of increased disclosure?
Poll
Loading …BrainTrust
Ken Lonyai
Consultant, Strategist, Tech Innovator, UX Evangelist
Ananda Chakravarty
Vice President, Research at IDC
I think it’s important to be transparent in whether you’re dealing with a person or a bot. I feel even when we were transparent in the past and made it clear it was a real person people didn’t believe us and instead of typing full sentences they would just send keywords assuming that would ping a response from the bot. It was very interesting. I think there is something to people being upset if they find out it’s a bot half-way into the conversation. There is nothing wrong with bot functionality, but in a perfect world I believe merchants should opt to offer both options.
Some companies may think they’re fooling customers with chatbots, but they aren’t. Those of us who find ourselves shouting “OPERATOR!” to get help from a human just don’t want to deal with an uncomprehending machine.
Retailers should be completely transparent and, more importantly, provide a way to get to a person quickly and easily. I’ve used a couple over the weekend which were terrible about this — had to go through their routine before I could get close to a human being. Which may emphasize that not only should retailers let customers know it’s a chatbot, but they should pay a lot of attention to the UX of that process.
I like transparency. Let the customer know they are interacting with a digital assistant — and let them know how to transfer to a live support person. It’s that simple. Inform the customer. If they are looking for basic information, such as shipping updates, it is easy to use the digital assistant. And if the customer does decide to talk to a live support person, the transition should be easy and seamless.
Although bots’ actions typically reveal their true identities very quickly, I feel like retailers would be best served by clearly labeling their agents as bots right up front, and combining that with an option to immediately reach out for a live agent. Perhaps an estimate of the wait time for a live agent would also help shoppers decide their course of action: see if they can navigate through the bot’s algorithms immediately or wait for a human being.
There should be 100% transparent and also AI incorporating Voice of the Customer should be employed, so the bots improve over time.
Transparency has always been a favorite of mine. Despite the cons to revealing that one is communicating with a virtual and not a real CSR, informing the consumer should always win out over shady practices.
The main objective of most consumers is to have their issues solved or their questions answered. If chatbots can do that effectively, then most people have no issue in dealing with them. However, chatbots are still incapable of dealing with every situation and human intervention is sometimes required. The skill here is to recognize these situations and quickly transfer to someone who can help, perhaps noting the transfer from bot to human.
Let me be perfectly clear. There is no personal connection here and it is very frustrating to your customer. Retailers should always answer with a REAL PERSON. Chatbots save retailers money but alienate customers. My personal least favorite sound is the fake key clicking noise that they use to simulate keystrokes. Most retailers have stores that cross multiple time zones that can spread the workload as they are not always dealing with peak floor times. This can translate to superior customer service and extra pay for associates.
It is critical to advise that customers are interacting with a machine — customers are able to work this out for themselves. Trying to mislead just undermines trust.
A similar approach appears on websites when viewing products or services with messages that inform the customer that someone else has just purchased this product. 90% of the time this information is false, it is easy to see this and it acts as distraction.
In today’s world of chaos and uncertainty, authenticity and transparency are the greatest gift brands and retailers can give to their customers and a key factor to business success. Shoppers understand the benefit of chatbots to retailers and can accept their use, but where things fall off the rails is when retailers try to pretend the bots are real. Shoppers are insulted by attempted deception and rarely fooled. It also more often than not leads to the perception of poor customer service. Retailers would be so much better served by being open about bot use, clear on its limitations, and offer a path to a live person should the chat bot conversation go off the rails.
Chatbots are a nice way to avoid a long hold if questions are straightforward. Identifying that the the bot is not a live person allows customers the option to not waste time with the chatbot first.
There will soon be a time when the sentiments will reverse. People will want to talk to a chatbot more than a human. That will happen when the AI technologies will improve further to be able to not just mimic human interaction better, but much more smartly and with greater speed. At all times, it is best to disclose and be transparent with customers.
Now that chatbots have been around for several years, most customers likely know how to work with them to get their points across. It is important to let customers know if they are dealing with a chatbot because then the customer can modify, then simplify their communication to get a better-desired result. Or, they may elect to postpone a sensitive or emotional conversation until a live person is ready. Best to make it fully transparent!
Consumers would be happier with getting a chatbot if chatbots were better able to serve their needs. Retailers should always be transparent with shoppers as to who they’re communicating with, but the root of this issue is about improving chatbots’ ability to solve shopper problems. As chatbots become more helpful, the transparency will actually benefit chatbot use because shoppers will realize they can actually get their needs met that way.
As a developer of enterprise-level chatbot/artificial assistant platforms, I like the idea of not disclosing that a bot is a bot to test the system’s capabilities to perform “like a human” and gauge user responses. However, my interests are secondary to user-centered design thinking, appropriate customer communications policy, and transparency, so of course, a bot should be disclosed to the user in the welcome message/disclaimer. In fact, it has been California law for a number of years now.
No one wants to be duped or feel duped. Not disclosing upfront creates that risk and the risk of alienating the customer. And … any organization implementing a bot that is expected to deliver real customer and corporate value has the bot platform backed by live agents to escalate to when a bot cannot resolve all issues. So it borders on outright disrespect to have a bot fake or heavily imply being a human that has to transfer a conversation to a “better trained” human.
For customers it doesn’t matter if the text on the other side comes from a human or not, as long as the problem is solved. If not solved, that’s when being a bot becomes an issue where customers will throw a fit about not having human customer service as well as the initial problem.
The smart way to engage is being transparent and setting up resources that allows a fast transfer to human support, with rapid identification that the customer needs more than just basic password reset help. Making a simple contract up front, letting the customer know they are dealing with a bot and will escalate to a human for any complex issue, allows collection of data functions and simpler tasks like password reset to be handled. This reduces resource time but leveraging the bot tech without alienating the customer and having them scream for a customer service number. Anything found in the FAQs can usually be handled by the bot- anything more complex requires human intervention.
Retailers should be totally transparent, but customers should be savvy enough to know the difference. It isn’t all that hard. I don’t see cons other than the fact that a lot of people want to talk to a human being, not an AI program. But those same people will be just as upset when they figure out that “Alice” is a bot.
Retailers, and all industries, should disclose when they are using a chat bot. This can be a very positive experience as people expect this in many cases, e.g. after a user has selected the type of problem they are trying to solve preference the next step with “Our automated system will now gather the information needed to simplify and speed the reset of your password.” To Cathy’s point, what is a bigger problem is not having an easy way to get to a human when it is needed. Transparency is best, e.g. “use our automated system or we can call you back in 90 minutes.”
Technology has enabled retailers and other companies to offer bots that can answer questions at any time of day or night. As the customer questions and interactions become more complex, the bots’ ability to satisfy disappoints. Knowing when to use bots and how to handoff to a live person is crucial. For those retailers that do have their associates perform virtual assistance, they should leverage that to the hilt to separate themselves from camouflaged ones.
What’s odd in this study is suggesting that we act like bots are real people and then reveal them later. No bot conversation should be implemented that way — we must be honest from the outset.
That said, customers are, in fact, quite willing to work with chat bots if we consider them the equivalent of a phone system attempting to guide the interaction to the most useful operator. There may be existing content which can answer questions. If so, customers will be pleased if they arrive there quickly and efficiently. If not, then get them to a human being fast.
Where the industry gets in trouble is attempting to make the chat bot feel human. Never lie to your customers.
Chatbots are a fantastic way for companies to increase the level of support they provide to consumers. By taking XX% of issues off of the plates of much more expensive humans, organizations are able to invest more heavily in providing top quality support when and where it is needed most. The newer consumers in the market both expect bots, and are becoming more accepting of them so there is no reason for companies to hide the fact that they are leveraging technology, as long as they make it clear that it is part of an overall commitment to deliver the highest degree of service possible!
I believe you can find success in a hybrid model while still being totally transparent. We offer links to our most commonly asked questions that are clearly prewritten and at the end show a Chat with Us link. We start the Chat with a bot asking for Email, Name and what their concerns are so we can link them to the most appropriate live representative. I like a bot for information gathering, but not as a way to dupe the customer into thinking they are conversing with a live person.
Generally, the more frustrating your issue, the less you want to speak to a bot. If merchants can work in logic where bots handle basic, more functional tasks and customers have the option to speak to a human for more pressing matters, then this could be a win/win for all.
However, some companies see the cost savings in automation with bots and over look the human need for compassion and empathy in more complex matters. Chat bots limited range of emotions can erode customer trust and brand favorability in critical moments and make customers feel like they don’t matter; just the opposite of what bots are intended to do. It’s a fine line but companies must disclose when “help” is a bot.
For now, it’s best to let customers decide which path they want to take. No one wants to hold for 22-minutes to get help but if a customer is willing to do it, then it’s clear that their issue is important and they expect to be treated as such.