November 8, 2024
©PhonlamaiPhoto via Canva.com
Will Consumers Warm Up to AI Customer Service Agents?
A survey from Salesforce found that nearly 75% of consumers want to know if they’re communicating with an AI agent, but 45% are also more likely to use an AI agent if there’s a “clear escalation path” to resolving a customer service issue.
The global survey of more than 15,000 consumers taken over July and August found that AI technology overall is raising the stakes around trust for brands, with 60% of consumers believing advances in AI make trust even more important. Nearly three-quarters (72%) trust companies less than they did a year ago, while 65% feel like companies are reckless with customer data.
Yet the survey shows some consumers are willing to work with AI agents if it more quickly resolves their problem:
- Over a third of consumers would work with an AI agent instead of a person to avoid repeating themselves.
- 30% of respondents — or 37% for Gen Z and millennials — would work with an AI agent instead of a person for faster service.
- A quarter of consumers — roughly one-third for Gen Z and millennials — would share their personal information with an AI agent so it can better anticipate their needs.
Salesforce wrote in a press release, “From alleviating clunky purchase experiences to difficult return processes, there’s an agent for that. But to build trusted customer relationships, brands need trusted AI agents that are grounded in transparency and the right data.”
The survey comes as Salesforce in September announced plans to roll out its Agentforce suite of AI-powered autonomous agents at Saks Fifth Avenue.
More pessimistically, a Gartner survey of nearly 6,000 U.S. consumers conducted in December 2023 found that 64% would prefer that companies didn’t use AI in their customer service. About half (53%) indicated they would consider switching to a competitor if they found out a company was going to use AI for customer service.
Their top concern about AI in customer service was that AI will make it more difficult to reach a live person, cited by 60%; followed by AI displacing people’s jobs, 46%; AI providing the wrong answers, 42%; their data being less secure, 34%; and AI biases not treating customers equally, 25%. The same survey found that only 14% of customer service and support issues are fully resolved in self-service.
“Once customers exhaust self-service options, they’re ready to reach out to a person,” said Keith McIntosh, senior principal of research, in the Gartner Customer Service & Support practice. “Many customers fear that GenAI will simply become another obstacle between them and an agent. The onus is on service and support leaders to show customers that AI can streamline the service experience.”
A survey of 1,000 U.S. adults taken in July and commissioned by the AI customer service company Cogito found that 39% of respondents believed talking with a chatbot or automated system is worse than talking with a real human, while 33% said it’s better.
Most respondents (77%) said that when engaging with customer service, they’d rather interact with a human using AI “in the background” than a fully automated chatbot. About a quarter (22%) were uncomfortable with a customer service agent using AI behind the scenes of a call in any way.
“While AI tools are seen as valuable for streamlining tasks behind the scenes, a clear preference exists for human customer service agents,” Cogito wrote in its analysis of the results. “Net-net: AI is best seen as a tool to empower agents, not replace them.”
BrainTrust
Carol Spieckerman
President, Spieckerman Retail
Lisa Goller
B2B Content Strategist
Sarah Pelton
Partner, Cambridge Retail Advisors
Recent Discussions
Discussion Questions
What role do you see AI playing in the customer service function?
Do you see the technology ultimately empowering or replacing customer service agents?
Poll
Loading …
If the AI service agents do a good job and help the customer with minimum hassle, then I don’t see people will have too many issues. However, the moment AI starts to fail by not answering questions properly, not showing empathy, giving poor responses, and so forth, customers will get frustrated. At that point – or for certain complex queries – it is important to have humans to assist. As I have said before, with AI is not either or – there is a place for humans as well as automation.
AI isn’t a substitute for human interaction (yet) and arguably makes human intervention more important than ever. Managing the hand-offs between high-tech and high-touch has emerged as a top challenge for any service-oriented business. I’m far from a curmudgeon and have recently been impressed with a couple of AI-powered customer service interactions that weren’t particularly straightforward. Even so, companies must remain vigilant and realistic about the limits of AI and the potential to do more harm than good when it comes to customer perception.
You’re right. I’ve been working with virtual agents for nearly 8 years and these findings are not novel. Consumers want transparency. Since the early days, more and more companies start engagement with a message confirming if the agent is human or AI. What is less common is transparency on how easy it is to escalate to a human (e.g. same as when you’re dealing with a retailer’s IVR and say “Agent” at any point). The findings in this article highlight how one of the most basic and least risky uses of AI still has not centralized around a common user experience.
Totally agree on the need for transparency and standards (if only informal). It’s kind of crazy that transparency is the exception rather than the rule given the mandate across other aspects of the business.
It’s up to the Agentic AI. If it’s effective, it’ll stay. If customers hate it, it’ll go.
Would you rather wait 10 minutes (or more) on hold to resolve a complaint or have it taken care of immediately? Most people would answer, “Immediately.” But to do that, they may have to use an AI-fueled digital solution. Today’s AI is different than just last year’s version, and it will continue to improve. The problem today is that not enough companies have adopted the latest versions of the technology, creating the typical consumer’s lack of confidence. AI can manage many simple/basic customer service functions. This leaves live agents available for more complicated issues.
For example, customers have adopted the concept of booking their airline reservations online—but it took a while. When customers have a problem, they pick up the phone and talk to a live agent. This has become acceptable to most of the typical airline’s passengers. The same experience will become the norm for customer service in many industries, especially retail.
Certain customer service chat experiences were so fast and smooth that I couldn’t tell if it was a human or bot helping me. If I get the answers I need fast, it doesn’t matter which one it is.
As AI evolves as a collaborative tool, it will also get smarter as customer service workers feed it new data on exceptions and nuanced queries. Over the long term, AI will replace many human workers.
AI agents for customer service is a best use case for AI. Lots of similar questions answered instantly and accurately without human training issues. Smaller, fast, low cost focused AI dataset. Add RAG to keep responses current. It’s a beautiful thing. We will look back and wonder why we put up with human agents guessing. Just attend to initial and ongoing training the AI.
Can you name 9me, please?
Walmart owned Myntra ecommerce platform in India uses AI to translate natural language-based inquiries into fashion recommendations. Sporting goods retailer Decathalon uses AI customer support to handle a much higher volume of customer service calls. HomeServe USA uses their conversational AI platform Charlie to quickly answer customer queries, initiate a claims process, and schedule appointments.
Asking if customers want to use AI probably generates a different response from asking if they want a faster, better customer service experience.
Well-trained AI agents are faster, more accurate and can route to live agents as required. There is risk of a poorly implemented AI experience.
Oh God…I just dealt with Walmart’s customer service bots. It was really, really awful
As long as the “ai” agents are as scripted as the outsourced agents, no, it’ll be a total waste of time. I’m always looking for a human. I was on with one last night who wouldn’t pass me along until she said the last line in her script (which was a meaningless platitude
Who is our example of an AI agent? Comcast? AT&T? Walmart? Do you know how badly they all suck?
Replacing. And with the replacement will come more frustration, more irritation, lower loyalty.
I hope the finance office enjoys the savings while they can because this will boomerang (sp) back to haunt them in the long run.
Amen
The best use of AI for customer service will be arming real front-line employees with additional information about their customers and products, and options to resolve situations. If it helps them be quicker, better and feel more empowered, it’s a winning opportunity.
Consumers will be fine with AI customer service provided that programs are set up will enough to truly resolve most types of consumer’s needs. But much like automatic phone systems, once the technology is unable to resolve a specific issue, consumers become very unhappy, at least until the option is made to talk to a human.
Most consumers are annoyed when “if this, press one,” and “if this press two, etc.” fails to offer a choice that fits their problem. We all have had frustrating experiences, where the system offers six or more “if this, press X” options, and still nothing hits the target.
But if consumers can still eventually speak to a human, or receive a call back in due time, all will be fine. Algorithms are pretty good, and many consumers prefer talking to a robot rather than a crabby human, or a human speaking with an accent that is hard to understand.
AI will probably take over a lot of routine tasks, but it’s more likely to support customer service agents than replace them. There’s a human touch—empathy, creativity, and complex problem-solving—that AI just can’t replicate right now. So, human agents will still be crucial, especially for handling sensitive, personalized, or high-stakes issues. AI is set to reshape customer service roles, allowing agents to focus on high-value interactions while it takes care of the repetitive stuff. In the long run, this partnership between AI and human agents can lead to more efficient and satisfying customer experiences.
Consumer experience with common chatbots (non-AI) has been pretty horrendous. Their algorithms suck. They can’t respond to plain-language queries. They send users to scripted content and generic help screens. We all perceive how they are designed to present instant (but usually useless) responses so that most users abandon their queries before a human being finally gets involved. When a live agent finally connects, they rely on the same mindless scripts. No wonder most surveyed consumers say they don’t want to interact with service bots.
There is potential for AI to make this better in two main ways, I think: One – train chat systems to understand queries from non-expert users and direct them more reliably to solutions. Two – make human agent systems smarter, so that they can stop relying on scripts and help customers move more directly to relevant solutions.
Consumer acceptance of AI bots is going to take some time. Enthusiasm will take much longer.
The old rules-based, scripted chatbots are on their way out. AI will usher in a new type of chatbot that likely will be far more helpful and empathetic (if trained well). I’d rather get my issue resolved immediately than wait around for a less helpful person who hasn’t been trained well. Check out bland.ai if you want to see the future of voice agents.
AI in customer service brings clear benefits, like faster resolutions, but it’s also raising valid concerns about transparency and accessibility. People don’t want AI to replace human agents; they want it to make things smoother without creating barriers to reaching a real person. Striking that balance will be key for brands that want to keep customer trust and make AI a positive part of the experience.
More Discussions
How Important Is ‘Vibe’ in Physical Retail?
Not attempting to be too playfully groovy about this question: According to a recent report from…Will Pinterest Translate to Shoppable TV?
Will Pinterest lead the way in shoppable TV?How Can Retailers Improve Shoppers’ Sense of Closure When Concerns Emerge?
How can retailers better provide closure and proper resolution to today’s shoppers?
If the AI service agents do a good job and help the customer with minimum hassle, then I don’t see people will have too many issues. However, the moment AI starts to fail by not answering questions properly, not showing empathy, giving poor responses, and so forth, customers will get frustrated. At that point – or for certain complex queries – it is important to have humans to assist. As I have said before, with AI is not either or – there is a place for humans as well as automation.
AI isn’t a substitute for human interaction (yet) and arguably makes human intervention more important than ever. Managing the hand-offs between high-tech and high-touch has emerged as a top challenge for any service-oriented business. I’m far from a curmudgeon and have recently been impressed with a couple of AI-powered customer service interactions that weren’t particularly straightforward. Even so, companies must remain vigilant and realistic about the limits of AI and the potential to do more harm than good when it comes to customer perception.
You’re right. I’ve been working with virtual agents for nearly 8 years and these findings are not novel. Consumers want transparency. Since the early days, more and more companies start engagement with a message confirming if the agent is human or AI. What is less common is transparency on how easy it is to escalate to a human (e.g. same as when you’re dealing with a retailer’s IVR and say “Agent” at any point). The findings in this article highlight how one of the most basic and least risky uses of AI still has not centralized around a common user experience.
Totally agree on the need for transparency and standards (if only informal). It’s kind of crazy that transparency is the exception rather than the rule given the mandate across other aspects of the business.
It’s up to the Agentic AI. If it’s effective, it’ll stay. If customers hate it, it’ll go.
Would you rather wait 10 minutes (or more) on hold to resolve a complaint or have it taken care of immediately? Most people would answer, “Immediately.” But to do that, they may have to use an AI-fueled digital solution. Today’s AI is different than just last year’s version, and it will continue to improve. The problem today is that not enough companies have adopted the latest versions of the technology, creating the typical consumer’s lack of confidence. AI can manage many simple/basic customer service functions. This leaves live agents available for more complicated issues.
For example, customers have adopted the concept of booking their airline reservations online—but it took a while. When customers have a problem, they pick up the phone and talk to a live agent. This has become acceptable to most of the typical airline’s passengers. The same experience will become the norm for customer service in many industries, especially retail.
Certain customer service chat experiences were so fast and smooth that I couldn’t tell if it was a human or bot helping me. If I get the answers I need fast, it doesn’t matter which one it is.
As AI evolves as a collaborative tool, it will also get smarter as customer service workers feed it new data on exceptions and nuanced queries. Over the long term, AI will replace many human workers.
AI agents for customer service is a best use case for AI. Lots of similar questions answered instantly and accurately without human training issues. Smaller, fast, low cost focused AI dataset. Add RAG to keep responses current. It’s a beautiful thing. We will look back and wonder why we put up with human agents guessing. Just attend to initial and ongoing training the AI.
Can you name 9me, please?
Walmart owned Myntra ecommerce platform in India uses AI to translate natural language-based inquiries into fashion recommendations. Sporting goods retailer Decathalon uses AI customer support to handle a much higher volume of customer service calls. HomeServe USA uses their conversational AI platform Charlie to quickly answer customer queries, initiate a claims process, and schedule appointments.
Asking if customers want to use AI probably generates a different response from asking if they want a faster, better customer service experience.
Well-trained AI agents are faster, more accurate and can route to live agents as required. There is risk of a poorly implemented AI experience.
Oh God…I just dealt with Walmart’s customer service bots. It was really, really awful
As long as the “ai” agents are as scripted as the outsourced agents, no, it’ll be a total waste of time. I’m always looking for a human. I was on with one last night who wouldn’t pass me along until she said the last line in her script (which was a meaningless platitude
Who is our example of an AI agent? Comcast? AT&T? Walmart? Do you know how badly they all suck?
Replacing. And with the replacement will come more frustration, more irritation, lower loyalty.
I hope the finance office enjoys the savings while they can because this will boomerang (sp) back to haunt them in the long run.
Amen
The best use of AI for customer service will be arming real front-line employees with additional information about their customers and products, and options to resolve situations. If it helps them be quicker, better and feel more empowered, it’s a winning opportunity.
Consumers will be fine with AI customer service provided that programs are set up will enough to truly resolve most types of consumer’s needs. But much like automatic phone systems, once the technology is unable to resolve a specific issue, consumers become very unhappy, at least until the option is made to talk to a human.
Most consumers are annoyed when “if this, press one,” and “if this press two, etc.” fails to offer a choice that fits their problem. We all have had frustrating experiences, where the system offers six or more “if this, press X” options, and still nothing hits the target.
But if consumers can still eventually speak to a human, or receive a call back in due time, all will be fine. Algorithms are pretty good, and many consumers prefer talking to a robot rather than a crabby human, or a human speaking with an accent that is hard to understand.
AI will probably take over a lot of routine tasks, but it’s more likely to support customer service agents than replace them. There’s a human touch—empathy, creativity, and complex problem-solving—that AI just can’t replicate right now. So, human agents will still be crucial, especially for handling sensitive, personalized, or high-stakes issues. AI is set to reshape customer service roles, allowing agents to focus on high-value interactions while it takes care of the repetitive stuff. In the long run, this partnership between AI and human agents can lead to more efficient and satisfying customer experiences.
Consumer experience with common chatbots (non-AI) has been pretty horrendous. Their algorithms suck. They can’t respond to plain-language queries. They send users to scripted content and generic help screens. We all perceive how they are designed to present instant (but usually useless) responses so that most users abandon their queries before a human being finally gets involved. When a live agent finally connects, they rely on the same mindless scripts. No wonder most surveyed consumers say they don’t want to interact with service bots.
There is potential for AI to make this better in two main ways, I think: One – train chat systems to understand queries from non-expert users and direct them more reliably to solutions. Two – make human agent systems smarter, so that they can stop relying on scripts and help customers move more directly to relevant solutions.
Consumer acceptance of AI bots is going to take some time. Enthusiasm will take much longer.
The old rules-based, scripted chatbots are on their way out. AI will usher in a new type of chatbot that likely will be far more helpful and empathetic (if trained well). I’d rather get my issue resolved immediately than wait around for a less helpful person who hasn’t been trained well. Check out bland.ai if you want to see the future of voice agents.
AI in customer service brings clear benefits, like faster resolutions, but it’s also raising valid concerns about transparency and accessibility. People don’t want AI to replace human agents; they want it to make things smoother without creating barriers to reaching a real person. Striking that balance will be key for brands that want to keep customer trust and make AI a positive part of the experience.