8 Conversational Design Mistakes You Can Learn From
⏱ Reading Time: 8 minutes
What is conversational design, and how does it differ from other types of setups? Think of it as a way to talk to your customers. You must engage and invest in them dynamically. They aren’t just reading an article — they’re interacting with it on some level. Conversational design typically involves methods such as live chat.
According to Gartner, about 100 million consumers shop via augmented reality (AR). People say they talk to bots almost as much as their spouses. If you aren’t looking at conversational design as at least part of your business website plan, you’re missing out on some opportunities to connect with leads.
However, there are some common mistakes people make when starting their conversational design journey. Engaging customers is about far more than just throwing up a live chat feature or offering customer service via SMS.
You must pay careful attention to the format, language and look of your conversational design. Keep an eye out for these errors, and your customers will thank you.
Conversational Design Mistakes
1. Creating Unnatural Interactions
Have you ever visited a website where the chatbot conversation made zero sense? With today’s machines understanding natural language patterns, there is no reason for a bot that doesn’t communicate well. Seek the best software and programming you can find.
Think about the way people talk. When they call and ask a question, what does the language sound like? Program those types of queries into your chat robot along with an answer that makes sense.
Have a plan in place for how you transition from the bot to a live agent. At what point does a real person step in and take over the conversation? You could use a bot to answer a list of 10 essential questions or collect information such as name and issue.
Some bots are more advanced than others, but they are still computers and won’t respond the way a human does. Decide what balance of bot and agent you’d like to implement.
Picture 1. Chatbot mistakes
2. Treating All Customers Alike
One of the things customers have come to expect in the last few years is a more personalized experience. Artificial intelligence (AI) has made some drastic advances. You can now program a computer to work as a personal shopper, offering suggestions based on past behavior and other factors.
When your site visitors feel like another number, they aren’t nearly as likely to buy. Fortunately, there are many ways you can personalize the experience with conversational design.
Instant interaction with a bot is just one method. You can also use text-based messaging or encourage interaction with videos, clickable elements and pop-ups that respond to specific actions.
3. Not Putting Safeguards in Place
One news report talked about a little girl ordering a dollhouse from her Alexa smart speaker. You’ve likely heard similar stories before. A conversational design that allows natural errors isn’t a very good one.
Spend time thinking through how people might accidentally click on features or place an order in error. Look for weaknesses in your conversational system. Hackers might exploit them, and customers might grow frustrated.
If you receive a complaint about your conversational design misfiring in some way, try to repeat the error and seek ways of fixing it for future clients.
You also need to consider how to secure the data customers share. They may engage in a chat where they give details such as account numbers, personal information and PINs. Although most third-party providers have security in place, cybercriminals can gain sensitive data.
Create policies about what information is shared online in the first place. Install firewalls and virus protection. Train your workers on how to protect their logins and customers’ data.
You can authenticate users in various ways. You can use voice and face recognition, passwords or third-party code generators.
Picture 2. Make sure you pay attention to online security
4. Writing Lengthy Messages
When creating automated messages, don’t make the mistake of writing a novel to answer a simple question. The very nature of conversational design is short and to the point.
According to Forrester’s research, live chat usage increased to 65% and is now preferred over email or phone support. Users see live chat as a fast, effective way of communicating with customer service.
Messages should be a few lines at most. Ask the questions you need to solve the problem for the customer. Tell them if you’re consulting a manager or looking something up, so they aren’t waiting long moments without a response.
Imagine the person is in front of you. If you had to ask your manager for input, you wouldn’t just ignore the customer. You’d say, “Excuse me for a minute. I have to ask my manager’s approval.” Treat online customer support with the same interaction they’d receive in a brick-and-mortar establishment.
Think about the way you send text messages. They are typically short one-liners, and your interaction on a live chat or SMS support should also be to the point. Too much information at one time can confuse customers and create aggravation with the process.
When programming automated answers, keep the same concept in mind. Responses should be as short but as informative as possible. If the response is longer, refer them to a FAQ or break it into multiple messages.
5. Creating Dead Ends
Since you’ll likely use artificial intelligence at least part of the time in your conversational designs, you want to make sure users have their questions answered. Walk through each of your automated responses and see where they lead.
A dead-end occurs when you didn’t fully answer a question, but the conversation concludes. Put measures in place to avoid this, such as always asking if the customer has additional questions. You can still route them to a live agent if the bot is out of responses.
Even your greeting could lead to lost interactions. For example, if you write, “Hello. Welcome to the site,” the customer may say hello and then leave. They won’t know what they should do next. Even generic “how can I help you” commands aren’t very workable.
Instead, go with a prompt, such as, “Would you like to know more about our return policy? You can ask me anything.” Giving the user a suggestion for the type of question sets the tone and clarifies how to move forward.
Look at your prompts as having a conversation with a nontalkative customer. You may have to drive the conversation forward by asking key questions. We’ve all been in situations where speaking to someone was like pulling teeth. Pretend that’s who you’re talking to when writing your prompts.
6. Making Your Design Intrusive
Most businesses still have a traditional website with conversational features. However, live chatbots can be pretty disruptive. If you’ve ever read an article and had a huge pop-up insisting they want to speak with you, you’ve experienced an annoying design.
The best conversational designs make it clear where the user must click to engage the agent, but they don’t intrude on the rest of the UX. It’s tempting to allow AI to push engagement, but it must occur organically.
Picture 3. Be careful about your design
7. Ignoring Mobile
According to Statista, 91% of total internet users access it via mobile devices at least part of the time. If your conversational design doesn’t factor in your smartphone users, you’re missing out on UX for at least half your audience.
Any type of conversational design should be thoroughly tested for mobile responsiveness. Pull up the interface on your smartphone and see how things work and if they look the way you want them to.
Fix any issues with tiny text or images sliding off the screen. Look for ways to engage mobile users, even if you must create a separate app.
8. Not Giving Your Bot a Persona
Your chatbot and other automated tools are the faces of your company. They may be the first interaction someone has with your brand, so make sure you give your bot a personality.
Just as you create a buyer persona to better understand your customers, make one for your bot. Does she or he have a sense of humor? Give the chatbot a name and have it introduce itself to your customers.
Your bot’s voice should match your brand’s voice. You likely already have a style guide for marketing purposes. How can you pull features from there to create a chatbot? Each response from your live chat should have a similar tone.
An excellent example of a company that offers a consistent customer experience is Chick-Fil-A. If the customer says, “Thank you,” they almost always respond with, “My pleasure.” The kind and welcoming tone is a voice and persona workers are trained in.
Customers may not be aware of the persona, but it offers a consistently positive experience no matter which restaurant they frequent.
Keep Tweaking Your Design
Don’t stop working on your design just because it’s great. There are always things you can do to further improve communication between your system and your audience.
Test new features, try different tactics and ask customers for feedback on their experience. Over time, your conversational design will far outshine your competitors’.
Eleanor Hecks is editor-in-chief at Designerly Magazine. Eleanor was the creative director and occasional blog writer at a prominent digital marketing agency before becoming her own boss in 2018. She lives in Philadelphia with her husband and dog, Bear.
Small Business Guide to Amazing Customer Support
More from our blog
Omnichannel analytics is the starting point if you’re trying to optimize customer experience. Here’s everything you need to know about it.
In-app messaging is great to keep your users engaged. Here are a couple of design tips that can increase its effectiveness.