Over the past several years, you’ve most likely noticed that chatbots have become a standard feature, particularly on corporate websites. Whether the user is a prospect on a buyer’s journey, searching for basic product info, or a customer looking to resolve a (hopefully!) quick service request, maybe an exchange or return, experimenting with the generative AI powered resource known as chatbot seems reasonable to many. After all, if it appears that you’ll be able to transact your business more quickly and easily, why not give it a try?

But unfortunately, many users have been disappointed by what they expected would be a good experience with a time-saving new technology. A survey conducted in 2023 found that while 68% of business website visitors have interacted with a chatbot just 35% agree that most of the time, the chatbot helped them achieve their mission. Meanwhile, 77% of users felt their chatbot experience was frustrating.

Duly noted but nevertheless, the corporate bosses are determined to making GenAI adoption a success. They’ve decided that chatbots are the future and the future is now; they envision chatbot as the portal where customers and prospects will access product information, register complaints and obtain refunds or exchanges—all without the assistance of the company’s (paid) employees, BTW.

The bosses are fully aware that these often mundane interactions are critical to customer satisfaction and loyalty, brand image and sales but as of now, users are reporting that their chatbot interactions are lacking. The 2023 survey referenced above also showed that 88% of business website visitors would rather speak with a human customer service agent via telephone when seeking information or resolving a problem, clearly indicating that today’s chatbot is in need of refinement. The current fixes don’t always work, according to users, resulting in what could be a great resource to come up short because the solutions that users want aren’t programmed into the thing. If the corporate champions of generative AI chatbots would like to realize its full value—and I suspect that even disappointed users would agree that the potential value of generative AI chatbot is considerable—the customer experience must be prioritized when programming and implementing the technology.

Customer service and IT teams are on notice to fix chatbot problems and fortunately, they are beginning to realize that the chatbot challenge isn’t primarily technological—it’s psychological. It seems obvious that chatbot developers would be greatly assisted by researching and verifying the solutions that users want and their expectations regarding the overall chatbot experience. In a 2024 survey, 85% of customer service leaders reported that they are currently exploring more user-friendly chatbot solutions.

So, Freelancer friend—does your website have a chatbot or have you, like me, hesitated to adopt a resource that ought to be a good thing, but your own user experience with the technology hasn’t lived up to your expectations? Or maybe those of you who have installed a website or social media chatbot might like to improve the experience? If you’d like to refine the responses, visit the Frequently Asked Questions document, which is the knowledge base for the chatbot. You’ll be able to make edits there and the chatbot will integrate the new information in its responses. Below are enhancements you may want to incorporate as you work to optimize your company’s chatbot experience and facilitate user satisfaction, enhance the perception of your brand and drive sales.

Conversational tone preferred

Research has shown that up to 35%  of customers behave differently when interacting with chatbots, as compared to a human agent, because interacting with a machine feels less personal. Make your chatbot communications more personal and welcoming by programming the system to respond like a friendly and efficient human customer service agent. A conversational yet professional tone, appropriate expressions of empathy (“I’m sorry to hear that”) and giving other common responses that (“How does that sound?”) can encourage users to feel comfortable and trust their chatbot interaction, behaviors that contribute to a positive customer experience. Also, creating a human-like avatar and giving it a name makes users feel that your chatbot is more relatable and trustworthy.

Your chatbot can even “learn” to give understated compliments that can discreetly recommend purchases. In fact, many customers respond favorably to gentle endorsement of their choices (“Yes, this tie will look both fashionable and professional”) and up to 12.5% are inclined to accept chatbot purchasing recommendations. Still, be aware that attempting to make your chatbot too chummy by programming in what can be perceived as false flattery and a turn-off. Overly personal language triggers suspicion and makes customers feel manipulated. Keep the compliments fact-based, professional and aligned with the professional tone that customers expect from a chatbot.

Customers in a hurry want just the facts

Customers who are in a hurry, or those who are upset and feeling that expectations have not been met, just want to fix their problem, quickly and satisfactorily. Feelings of disappointment and frustration override the preference toward human-like chatbot responses. Research has found that dissatisfied customers were 23% less satisfied when met with conversational chatbot “empathy” when there is a problem to fix. At those times, customers prefer a straightforward, all-business chatbot interaction that is fast and efficient.

Furthermore, conversational responses can raise expectations and inadvertently heighten frustration if chatbot responses do not quickly and accurately produce problem resolution. It will be useful to program your chatbot to shift into all-business mode and respond quickly and clearly when frustration or urgency are detected. In high-stress situations, it’s better to not seem “human” at all.

Advertise accuracy

Advertising the potential benefits of using your chatbot is a good thing! Let visitors to your website or social media sites know that your chatbot is able to quickly and correctly answer questions and solve problems, save time and make accessing information easy. Broadcasting the likelihood of success generates the expectation of a positive experience, promotes trust in the chatbot and encourages customers and prospects to take a chance with your chatbot. Inviting website visitors to “try it, you’ll like it” can potentially increase chatbot engagement by up to 22%, according to a 2021 study.

This point was reinforced in another 2023 study of chatbot interactions that revealed when users engaged in identical conversations with human customer service agents and the chatbot, 8.5% reported they were less satisfied after using the chatbot. But when chatbot benefits were highlighted, for example, announcing 24/7 service availability, 37% of that same subject group reported they were more satisfied with the chatbot than with the slower, and less-available, human agents. Give your chatbot an enthusiastic endorsement and you’ll encourage its use.

Getting better all the time

The best way to improve the performance and acceptance of your chatbot is to help it get “smarter”. Research shows that users were up to 17% more likely to accept the chatbot’s suggestions when told that the chatbot is on a continuous learning curve and is not limited by a static algorithm that’s upgraded only intermittently. It seems a chatbot that’s “enrolled” in ongoing continuing education is favored over a presumably knowledgeable human customer service agent.

In other words, tell the world that your chatbot is always being improved with the ongoing addition of new information and user feedback. It’s like working with an enthusiastic young intern—users may be inclined to forgive occasional small errors and appreciate that each interaction will be better than the last. Without this explanation, some users might have limited confidence in this resource (perhaps a result of underwhelming, if not disappointing, chatbot experiences on other sites). Make the effort to optimize how your chatbot communicates by training the technology to deliver customer service that meets, or exceeds, user expectations. Finally, know that it is also important that human interaction should be available to your customers and prospects when they request it.

Thanks for reading,

Kim

Image: © Dzmitry Auramchik