©NicoElNino via Canva.com
Are AI-Driven Chatbots Only Frustrating the Online Shopper Journey?
AI-powered chatbots were seen as the most disruptive artificial intelligence (AI) application causing friction in online shopping experiences, according to a recent survey of 1,000 U.K. consumers from tech provider Intellias.
Of the survey respondents, 25% agreed chatbots are the area where AI most disrupts their shopping experiences, and 23% said that interactions with AI chatbots were the stage in their online buying journeys where they were most likely to abandon a purchase.
The findings follow separate research from last fall by Talkdesk, which found that if given inaccurate recommendations from an AI-supported device, 68% of U.S. shoppers would lose trust in brands, 65% would not positively recommend the brand to loved ones, and 53% would not shop again at that brand.
The Intellias’ poll found that 49% of survey respondents don’t mind that retailers use AI as long as the experience isn’t “clunky.” Two-thirds don’t mind when retailers use AI if it is to automate repetitive or monotonous tasks as long as it does not replace human interaction.
Almost three-quarters (74%) agreed that “a blended experience of both automation delivered through AI and human interaction would always be needed in retail, regardless of how good the AI technology becomes in the future.”
AI-supported chatbots have received more attention with the arrival of ChatGPT in November 2022 and the rollout of several GenAI shopping chatbots with conversation capabilities from Amazon, Google, Instacart, Mercari, Carrefour, Kering, and others.
A global survey as part of IBM’s study, “2024 Consumer Study: Revolutionize retail with AI everywhere,” found the primary reasons adults are interested in using AI as part of the shopping journey were to research products or get information, cited by 86%; get services/answers or resolve issues, 82%; look for deals and promotions, 79%; look for reviews or recommendations, 78%; and shop for products or services, 77%.
Some early reviews of GenAI-enabled shopping chatbots haven’t been promising, with a few pointing to shortfalls in the technology’s promised ability to refine and personalize searches as well as to remember past searches or purchase history.
In a review of Amazon’s new Rufus shopping assistant, Shira Ovide, technology reporter for The Washington Post, wrote, “In general, I thought the shopping bot was at best a slight upgrade on searching Amazon, Google or news articles for product recommendations.”
Discussion Questions
Are AI-supported chatbots generally causing more consternation than solving problems for online shoppers?
Have you had an encouraging experience with GenAI-supported shopping chatbots?
Poll
Loading …BrainTrust
Raj B. Shroff
Founder & Principal, PINE
Lucille DeHart
Principal, MKT Marketing Services/Columbus Consulting
AI chatbots have a role to play. Many are good at answering basic queries and questions, such as when items will be delivered, and doing so efficiently. That’s good for reducing labor costs and provides a reasonable degree of customer satisfaction. Where they fall down is with more complex questions, with nuanced questions, or in areas where some empathy and reassurance is needed. There they can be a complete pain and result in a lot more friction for the customer. The answer, as always, is to have both AI chatbots and human assistance, and ensure that consumers have a choice over how they interact.
Chatbots still largely suck, in my personal experience, but this will change. Notwithstanding the tepid early reviews of advanced systems like Rufus, they will all get better. And I suspect it will happen relatedly fast, given how rapid Gen AI in general is advancing. Consumer expectations have been raised by the AI hype cycle, and so now these systems have to start delivering the user experience that they claim they will – consumers are ready for great chatbots!
Never underestimate the power of ingenuity. Chatbots/AI will continue to perfect their interaction with humans and will do so with infinite information at the ready. From product details to competitive comparisons, AI will streamline the shopping experience. One day consumers will look back and try and remember a time when they didn’t use AI and wonder how they managed.
Chatbots have the potential to provide the most accurate information. They only use data they are trained on. No opinion. No guessing. No pretending to know. A bit more training and time needed to refine. Those experimenting early might take a few hits now, but they’ll capture the data and experience to be ahead faster.
This class of chatbots is still in its infancy so at this point it’s hard to judge. Speaking as a consumer I have been both helped and frustrated by AI-driven chatbots, but I can say the same about speaking with humans online. The media is all over the AI coolness factor but I think we have to give it time to develop.
So right. We haven’t even imagined where AI development will take retail.
Ha yes! Human interaction can be very variable these days!
I have never had a positive experience with a Chatbot. In any use case. Ever. While I agree this technology has promise, I believe it is way off in the distance before it becomes useful.
We’re still in the infancy of AI chatbots and though some of the interactions are kluge and frustrating, they will improve and enhance the overall shopping experience. In 12-24 months, you’ll see a dramatic uptick in the way chatbots will perform, especially with general topics and information exchange. But don’t think they’ll run your business. The art of retailing is best left with the human mind and customers love to be exposed to this unique and differentiated thinking that no bot can offer.
If the host site for the chatbots has it well integrated, it should not cause consternation. AI-supported chatbots are easily able to handle the most simple queries and mundane tasks. This allows the human support staff to focus on the more complex questions and work 1:1 with shoppers should the need arise. Like everything, someone always finds problems with newer technology. However, over time, this will really help make ecommerce and support a way better experience.
I have had frustrations with chatbots when the shift to a human was too difficult or took too long. But I’ve also had many great experiences in which chatbots answered an easy question really fast and I could continue on shopping or ordering.
As retailers implement AI on their shopping sites, someone says it can do this or that. For me, I don’t care if AI is helping me or not. Just get out of the way and make my purchase smaokoth and simple.
Ah, the sound of the other shoe dropping. Isn’t this the same tech that is going to replace humanity? Maybe we should wait just a little longer before replacing those millions of customer service reps.
Dismiss the hype, not the technology. The technology is awesome but it wasn’t ready to live up to the hype.
Chatbots are mostly annoying and have remained annoying because they’re not human and neither the intro of NLP (natural language processing) or AI (which NLP claimed early on) have moved the needle much, YET. But they will.
Hype is the path to the dark side; patience, Danielson.
Chatbots will improve and become increasingly helpful with the incorporation of AI. Old school chatbots really weren’t all that helpful. Now, with AI, and training the chatbot on all relevant material, a much more intelligent and conversational chatbot will do much better. Just because these weren’t all that helpful in the past doesn’t mean they won’t be in the future!
There is great potential for better AI chatbots, because let’s face it, they are largely frustrating customers today, myself included. As LLMs are applied they will hopefully be able to pick up content and context that will result in being able to either resolve customer issues and concerns or pass the customer to a live person who can. Keeping my fingers crossed!
On the one hand, unless a Retailer is going to train an LLM with proprietary data that is not available in the public domain, (and that’s a pretty heavy lift) there is nothing a chatbot can answer that a customer can’t self-serve themselves with a Chat-GPT. On the other hand, perhaps a chat bot, on a clear static free line, that is audible and understandable will be an improvement over some of the call center experiences present today.
My larger concern will be the diminished access to a real human being with decision making capabilities and a focus on customer service. Today – if I have a problem with a phone or cable bill or service, or a health care claim I won’t try to resolve it unless I have at least a 2-hour window free. If you think it’s hard to talk to an informed decision maker today – you ain’t seen nothing yet.
You raise a good point – making it difficult for customers to cancel service or migrate to cheaper plans seems to be the current strategy for cable and health insurance, and taking humans with their organic empathy out of the loop is a great way to prevent anyone from finding clever billing codes or making end-runs around the system.
Anything related to artificial intelligence and GenAI is in its infancy stages and will continue to evolve over time to deliver value and drive improving associate and customer experiences. Increasingly, we have seen that with call centers and customer service needs, there is an aggressive push for chatbots and AI-driven capabilities to handle a majority of the fundamental questions consumers may have. There are clear labor and cost savings attributed to this operating model shift.
However, as many have experienced, the AI chatbot model is quite fragmented and continuously improving to drive more value. There is a degree of frustration as consumers may not have their needs addressed in a timely manner as they would with a human-to-human interaction.
However, we should expect an acceleration in the retail chatbot race, especially as Amazon recently released Rufus, a new generative AI-powered conversational shopping experience within its mobile commerce platform. There is plenty of room for improvement, but this is a typical part of the innovation process.
In my opinion, AI-supported chatbots often create more frustration than assistance for online shoppers. While they have potential benefits, like automating tasks and providing information, many users find them clunky and unreliable. In fact, in many instances, the AI chatbots fail to understand queries and provide irrelevant recommendations. As for GenAI-supported shopping chatbots, while some have shown promise in assisting with basic tasks, others have fallen short in personalization and responsiveness. Overall, there’s a clear need for improvement to ensure that AI chatbots truly enhance the online shopping experience for consumers.
Shoppers want AI chatbots to give them an advantage in finding deals and getting faster customer service when things go wrong. They are long past the “wow, it’s a robot!” phase. Retailers need to be more strategic in their use of AI chatbots and spend more time designing processes to be faster, more efficient, and better aligned with their brand.
Issues like inaccurate recommendations and a lack of personalization can cause frustration and worse: they erode trust in the brand. That’s why retailers need to use chatbots to enhance, rather than hinder, the online shopping experience. With advancements in AI algorithms and strategic design, AI chatbots can evolve to meet consumer expectations and even provide a competitive edge.