How Retailers Can Avoid Generative AI Failures in Customer Service
Generative artificial intelligence (AI) has become the technology to watch, and major retailers are starting to leverage this breakthrough new tech. Insider Intelligence reports that Walmart is using generative AI for its “Text to Shop,” which lets customers add items to their shopping cart by texting or speaking the item they’d like to purchase. Instacart is also experimenting with generative AI to create shopping lists with their “Ask Instacart” feature.
However, the biggest AI game changer for retailers will be in customer service—a $500 billion market and growing.
Generative AI for Customer Service
By 2027, Gartner predicts chatbots will become a primary customer service channel. Chatbots have been around for years, but remain unpopular with customers. Generative AI can reimagine how we experience chatbots, so that they work seamlessly across channels, languages and time zones.
Generative AI will allow your bot managers and conversational designers to focus on editing, fact-checking, and reinforcing your brand—instead of churning out FAQ content.
Of course, this potential comes with valid concerns about AI safety, like inaccuracy, irrelevancy and hallucinations (when AI produces an answer that sounds credible, but isn’t factual).
Here are some ways to proactively keep your brand reputation intact.
Strike a Balance Between Innovation and Safety
Retailers want to keep pace with technological innovation, but never at the expense of brand safety. Generative AI can be a CX superpower, but not if it goes off the rails or delivers false information.
The generative AI technology you employ should be able to read conversations between AI and customers and understand —with better than human accuracy —whether or not a conversation was relevant, accurate, and safe.
If your chatbots aren’t rendering correct and relevant answers, it’s time to reexamine. Don’t slam the brakes on AI-powered customer service altogether —once you get the right content in place, you can take your hand off the wheel .
Make Automated Resolution (AR) the New Gold Standard
I truly believe that Automated Resolution (AR) should be the north star metric of any CX organization.
AR is a fully automated interaction between a customer and a business that accurately measures whether a customer inquiry was resolved without human assistance. Historically, customer service measured whether a conversation was contained to the chatbot without needing an agent, but ‘containment’ doesn’t demonstrate if the inquiry was resolved.
By measuring AR, your customer and your business are aligned to the same goal: quickly solving customer issues without human intervention. Of course, none of this is static —you need built-in, continuous, iterative testing and customer-ready quality standards.
Check Your Policies Documentation and API Strategy
AI generated responses should be limited to your policies and company documentation, and not the wider internet. For example, customers shouldn’t be able to engage chatbots in off-topic conversations that they can later screenshot and use to damage your reputation. Generative AI may be capable of commenting on political and social debates— but your customer service bots shouldn’t.
The answers generated should be grounded only in your company knowledge base. Be sure to conduct regular audits to ensure documentation is up-to-date, clear and thorough.
Your API strategy also matters, because your systems need to speak to each other to access order numbers, trigger returns or reset passwords.
Zero In On Your Brand’s Chatbot Persona
Think carefully about your bot persona and how it matches your brand. Whether that persona is playful or witty, an engaging chatbot personality will help bridge the gap between humans and bots to boost brand trust.
We’re still in the early days of generative AI, but we know its enormous potential when trained and implemented in the right way. As retailers navigate this new frontier, mitigating risks will be key to preserving brand equity.
Mike Murchison is CEO and co-founder of Ada, an AI chatbot and customer experience platform.
Mike Murchison is CEO and Co-founder of Ada, the world’s leading customer service automation company. Since 2016, Ada’s AI-powered platform has automated more than 4.7 billion customer interactions for brands like Meta, Verizon, AirAsia, Yeti, and Square.