Mike Myer, CEO and founder of Quiq, the AI-driven chatbot solution for customer service, whose clients include Nespresso, Wella and Spirit Airlines, believes that generative AI promises to change our world in the way the internet did back in the mid-1990s
And, just as the internet took time to evolve from those early CompuServe and dial-in modem days, so it will take years before we fully comprehend how the world has changed through generative AI, he says.
Quiq has been using so-called conversational AI technology for customer service chatbot questions in sectors including travel, hospitality, retail and e-commerce.
Its asynchronous chatbot technology uses AI so that you can break off your conversation – none of that agonised waiting wondering if the chatbot or human agent has disappeared – close your laptop and pick up the conversation hours later. With notable results: Quiq claims 80 per cent of inbound requests resolved and a decreased call volume by 20 per cent.
We sat down with Myer to discuss the implications of generational AI for customer service and the wider implications for us all.
What is the difference between generative AI and conversational AI?
At this point in time, some of the underlying algorithms that get used in conversational AI are generative in nature. The underlying question comes down to, is ChatGPT going to replace customer service at large brands? ChatGPT is absolutely amazing at answering questions. From a customer service perspective, there are two fundamental concerns about what chat GPT delivers when it’s applied to the customer service use case.
First, when you’re giving customers advice, you want to be as accurate as possible. There’s an underlying risk that ChatGPT can result in hallucinations, and you’re going to be off brand. And the more that somebody messes with ChatGPT, the more off-brand it potentially could become. From a reputation standpoint, the folks who run customer experience want to be absolutely certain that they’re giving the best possible answers. ChatGPT represents a roulette wheel: most of the time, it will come out on black, and you’ll be good. But every now and then the ball will drop on red, which could be, uh-oh, embarrassing.
Second, are the large language models. They are built to consume all the information on the internet. But it’s not very often when somebody contacts customer service, they’re asking question which can just be solved on the internet, because if a person could find the answer on the internet, they would have done so already. They’re calling because they need something changed on their account, and that data is internal to the company. Their account information. ChatGPT or generative AI more generally, could be used to improve an agent’s performance or to help an agent write a better response. But the actual data and the answers which the agent needs to provide in customer service, ChatGPT is not going to have the that access to that information.
Although generative AI will have a role when it comes to customer service, it won’t be the dominant technology?
It’s not a replacement, it’s a performance and efficiency improver for the customer service agent. For example, a customer contacts a company because they are unhappy with something they’ve bought and want a refund. The agent needs to apply some human logic as to whether or not the customer qualifies for the refund, which means looking into their account, which again is internal data inaccessible to a public generative AI algorithm.
Where generative AI would have a really strong use case is helping to craft the response, so that it’s not too terse or long or whatever. Instead of the agent spending time composing the message to the customer, generative AI can produce the message – but it’s up to the agent to decide on the meat of what’s being said. Generative AI can do the flowery language around it.
Over the past few weeks, we’ve seen a gazillion articles and think pieces about generative AI. Is it just a hype cycle and, if it is, will it just fade away?
I don’t believe so. I think generative AI will fundamentally change some of the mundane things which have to do with either generating language or summarising blocks of language. It’s not going to replace traditional business functions, it’s going to make them better. I can’t think of another piece of business software that could be so revolutionary. Frankly, it’s freaking amazing.
That said, it’s still an incremental step – sure, it’s a bigger incremental step – but it’s not a replacement step.
At one end of the spectrum, we have Chicken Littles saying, the sky is falling on our heads, millions of people are going to be thrown out of work, and, at the other end, kind of a meh shrug.
Well, I can only answer this question relative to helping customers through customer service. On the Chicken Little side, you’ve got people saying, generative AI will completely replace the customer service function. The other is that it’s not going to have any impact at all. I’m probably 60 per cent it’s going to have a fundamental impact on the customer experience as we know it.
Looking into your crystal ball, how sophisticated are chatbots going to be in five years’ time? At the moment, you get three boring questions from a chatbot and then ask to speak to a human.
We’ll see some kind of connecting up between generative AI and proprietary back-end data. If you say, “I’m travelling to Phoenix today from San Diego airport, what’s my gate?”, generative AI at the moment isn’t necessarily appropriate to build a chatbot to answer that question, because you need like the actual schedule information in real time. In three- or five-years’ time, you could have multiple layers, where you inject real-time information in front of the model.
How important is generative AI as a tech development?
This is probably the most transformational moment in the past ten years. The internet in the late 1990s and 2000s created an enormous change in the way the world worked. And I think we’re at the moment of a second transformation in our lifetime of how the world is, which will take years – just as the internet took years to transform into what it is today. Generational AI is the start of another huge transformation.
I’m not Chicken Little in like, oh my gosh, this is going to be the end of humanity; I’m much more, this is going to be an enabler which helps people get stuff done. It’s not going to replace people, especially in the customer experience world, it’s just going to make the experience better for the consumer and reduce costs.
People are not going to have to do menial routine jobs, they’re going to be able to do more powerful jobs in the future because they’ve been empowered by generational AI doing those menial tasks for them.
More on generative AI
We chat with ChatGPT itself about the future of AI – What can ChatGPT tell us itself about the future of AI? What are the best use cases for Generative AI? And will artificial intelligence one day surpass humanity, leaving us behind?
The challenge of using ChatGPT for search engines – Large language models (LLMs) such as ChatGPT may be emerging as complements for search engines, but there are still pitfalls to consider
Will ChatGPT make low-code obsolete? – Romy Hughes thinks that ChatGPT could do what low-code has been trying to achieve for years – putting software development into the hands of users