Chatbots are everywhere. These “conversational agents” can help us book a flight, order a pizza, check our bank account balance, confirm an order status, and more. And many of us don’t care one whit that we’re relying on a software application for help. However, it appears that a good percentage of the population doesn’t like chatbots.

According to research by Forrester, 54 percent of U.S. online consumers believe that interacting with a chatbot will have a negative impact on their quality of life. “Consumers anticipate the worst when engaging with a chatbot today and haven’t yet seen a chatbot that can meet or even exceed those expectations,” reports Forrester analyst Ian Jaocbs.

That’s a pretty grim review. So, what’s the problem? Well … chatbots aren’t human. They can’t (yet) contextualize information, which is core to natural human dialogue. For the most part, chatbots are great at answering simple questions or handling routine tasks. But when a customer seeks more complex assistance — and maybe a little reasoning or empathy — things can get bumpy quick.

Good News: EQ is Already Out There in Chatbot Land

Research from PwC found that 71 percent of U.S. consumers would rather interact with a human than a chatbot or some other automated process. Obviously, that’s not great news for the 80 percent of businesses that already use chatbots, or plan to do so by 2020, according to research from Invoca and Adobe, which is featured in the report, Emotions Win: What Customers Expect in the Age of AI.

But all is not lost, provided that these companies keep their brands’ emotional quotient (EQ) in focus as they seek to further technify their customer service operations. EQ, as explained in our Emotions Win report, is the ability to recognize and respond to a person’s emotional state. Among the top EQ qualities that consumers say a customer service representative, AI-powered or not, can bring to the table are problem-solving, efficiency and adaptability.

Only 22 percent of consumers surveyed for the Emotions Win report said that chatbots have the edge over human hearts in delivering the best EQ—but that can change. Leading companies are already making strides in creating engaging chatbot experiences that are designed to give their customers the feeling that they’re making an (almost) human connection with their brand. Here are just a few examples:

Amtrak

Amtrak’s virtual assistant “Julie” is the digital-age version of the rail services provider’s original telephone-based customer service agent of the same name. Julie helps users navigate Amtrak’s website, and can answer questions based on natural language and sever up links to other useful resources on the site. Julie answers about 5 million questions annually—and she must be doing something right because Amtrak has seen an 8X return on its investment and gets 25 percent more bookings with help from the bot.

Casper

Casper’s Insomnobot-3000 is “a friendly, easily distracted bot designed to keep you company when you just can’t fall asleep.” The mattress company introduced the bot a couple of years ago. Insomnobot-3000 doesn’t sell mattresses, but it will engage in simple, lighthearted banter in the wee hours of the night. And while some people think the bot is a bit weird, there are no doubt plenty of insomniacs who find it comforting to have “someone” to commiserate with via text late at night when counting sheep isn’t cutting it.

H&M

Who doesn’t want a personal shopper? This global fashion company’s chatbot helps mobile customers find outfits based on information such as their gender and style preferences. If a customer doesn’t like an outfit suggestion, H&M’s chatbot (which also smoothly peppers its text dialogue with emojis and hipster vocab like “Coolio!”) will offer up another suggestion. And when a customer is ready to buy, the bot guides them through that process, too.

The Future is Human and Automated Support

Planning a trip by rail, coping with insomnia and finding the right outfit can all be stressful experiences for consumers, to varying degrees. Chatbots, clearly, can help to navigate these challenges. But when consumers are faced with making a big or complex purchase (think: car, house, major appliance, health insurance, etc.), their need for EQ in their interaction with a brand increases, according to the Emotions Win report.

Given that finding, it’s not surprising to learn that more than half of the consumers (54 percent) surveyed for the report said they believe the future of AI for customer support should involve a combination of human and automated interactions. The question is, how can businesses ensure their chatbots will meet their customers’ high expectations today and in the future? Here are a few strategies that can help:

  • Experiment internally first: Forrester’s Jacobs offers this advice in his blog post about “chatbot hype”: “Start by turning your chatbot inward — testing out new capabilities on your agents is a great way to see what works while boosting employee performance. … By taking it slow and steady, you’ll ensure that your customer service chatbot isn’t causing more problems than it’s resolving.”
  • Use data to personalize interactions: Make sure whatever platform your chatbot uses can integrate with key systems that can pull in relevant customer information. Data plays a vital role in building a foundation that demonstrates EQ and drives sales. (And remember, your human agents need access to data insights to provide great support, too.)
  • Treat your chatbot like a person: Your customers won’t feel like they’re in a humanlike interaction with a nameless bot that has no personality. For inspiration in creating a memorable and engaging chatbot persona to represent your brand, look to companies like LEGO. It’s Facebook chatbot messenger “Ralph” helps consumers during the holiday season to find the perfect LEGO gift. The “gift bot” character brims with personality and is all about being fun and helpful.

Chatbots need to evolve—and they are. But even the most sophisticated, AI-powered bots have their limitations when it comes to providing the EQ consumers crave. So, a final tip for chatbot success is to set expectations with your customers up front. Being clear with customers about what your chatbot can do, and helping them to have an efficient and productive interaction, can help them to have a more positive chatbot experience overall. Knowing the parameters of your chatbot’s capabilities can also help a customer decide whether they need to talk to a human representative, instead. Just be sure to make that process easy and fulfilling for them, too.