The Pros and Cons of Chatbots and How to Use Them to Gather Rich Data

Chatbot technology and user experience have come a long way since the days of Microsoft’s Office Assistant named Clippy. When designed well, today’s chatbots can help businesses automate, assist with user experience, and gather rich data from its customers.

Overlap’s design team uses chatbots for certain types of design research because they can access user data in a way surveys or person-to-person interviews cannot. Chatbots can produce phenomenal results, but to get there, they take plenty of planning, thoughtful design, and continual monitoring.

Overlap’s Senior Designer Dave Dowhaniuk participated in a Twitter chat with Startups Canada that dug into the pros and cons of chatbots. You can view the full chat on Storify and continue reading for Dave’s expanded answers on taking a human-centred approach to chatbot design.

Q1) What is a Chatbot? How and why are entrepreneurs using them?

A chatbot is an application used to simulate a conversation with a human, usually over the internet. Chatbots are being used to benefit startups by filling frontline customer service or technical support roles. They can even book restaurant tables or teach kids about dinosaurs.

Q2) What services do chatbots provide for startups?

Many organizations are replacing their first round of customer service or support with chatbots.

At Overlap, we use chatbots for design research. It’s an engaging and human-centred way to collect both qualitative and quantitative data. By using our understanding of conversation and subject matter, we are able to use chatbots to engage more like we do in our in-person interviews and engagements.

With good question and answer piping, a chatbot can determine human needs and keep people entertained, even if a human is on the way. By providing your customers with information, you can also collect deeper information about their questions, comments, difficulties, and concerns. You’re creating a cycle of customer experience improvement.

Q3) What are the steps to implementing a chatbot?

Start by making sure a chatbot is right for you and your customers. If it is, then you can think about what you need your chatbot to do.

What problem is it solving?

How do clients prefer to be interacted with?

Do they want a lot or very little personality?

Next, you can work out your chat flow (we do this on Post-its). Consider all possible questions and how different answers could change the desired effect of the chat.

Form your questions by looking at product reviews, speaking with frontline customer service reps, and co-designing your questions and answers with real customers. We determine our questions and answers by talking to people and kids. We listen to everyone as humans, and it’s that step that produces an exceptional chatbot.

After all of the above, you can think about programming.

Q4) How does a chatbot help organizations understand what their client demands are?

It’s all in the design process. You need to make sure you’re designing a satisfying customer experience while being efficient. You have to put effort into the design. Otherwise, you’re asking everyone the same question and not personalizing it, like a traditional survey.

A great example is when Co-designing with kids. You need to first uncover the way kids prefer to be talked to and asked questions. It’s about the human needs too.

Through good chat flow design, you can begin to pool data and use the responses your chatbot didn’t quite understand. You can build a database of client wants, needs, and demands so you can continue to build and refine your chatbot, customer experience, and product.

Q5) Chatbots can be great for customer service. How can entrepreneurs make sure they aren’t giving off a robot effect?

Who says you don’t want to give off a robot effect?

We chose a Chatbot for our work co-designing with children because we found that people (especially kids) are more comfortable opening up to a robot than a human. This is invaluable when you are surveying kids about personal or sensitive topics.

We also use the co-design process as a way to teach media literacy to kids. They learn how easy it is to create a simulated conversation while learning important STEAM (STEM + the Arts) skills. This is where the importance of the Arts in technology is made highly apparent. It’s how you learn to make products for customers not yourself.

Q6) What are the cons or potential risks associated with chatbots? Is it a case of ‘you get what you pay for’?

Many people, myself included, still prefer the human touch. The greatest risk for chatbots is not maintaining a human-centred approach throughout the design.

Remember that you want to improve customer experience and not detract from it. There are plenty of poorly thought out chatbots, including a certain bouncing paperclip that often comes to mind. It can seem fun to give a lot of personality to your chatbot, but remember your client’s needs come before your creative fun.

More on What Made Clippy, the Microsoft Office Assistant, So Detestable?

Q7) How can entrepreneurs manage their chatbots? Is it a ‘set it and forget it’ program or do you need to analyze responses?

I would never recommend a “set it and forget it” mentality when it comes to any part of your business. You can use chatbots to collect customer insights too. Learn from your chatbot and your customers.

Don’t just jump on the trend! Truly understand the power of AI and chatbots, and the purpose you’ll use it for.

Q8) From start to finish, how much do chatbots cost for Canadian Entrepreneurs?

We are currently programming our chatbot with a node-based programming application called Motion AI.

It’s a simple and intuitive interface that doesn’t require code. The product is rapidly growing, and it has a very responsive Slack channel for quick questions and feedback.

Prices range from $15 a month to $1000 a month for up to 250 bots, but there’s also a free prototyping version to get you started.

Q9) Can chatbots assist with language barriers? How?

With a well-designed chat flow and a great translator, you can breach language barriers.

You can also write your chat flows in a simple way that’s easy to understand no matter what your language level. This is perfect for children or for interacting with someone at a lower literacy level.

Q10) What are other ways artificial intelligence can help organizations automate and provide quality customer service?

Just using a chatbot or AI will not automatically provide quality customer service. Any AI still requires planning, good design, and an understanding of your customer’s needs.

Make sure you’re not just thinking about the bottom line and don’t use a chatbot as a solution to all of your customer service needs.

Q11) Where can entrepreneurs go today to access more information and to setup a chatbot for their organization?

There are plenty of blogs posted by FastCompany, Chatbots Magazine, and others that continue the conversation about the pros and cons of chatbots.

Here are some articles to get you started:

The Complete Beginner’s Guide To Chatbots

Essential Tips to Keep in Mind When Building a Chatbot

Is it OK for a Chatbot to Impersonate a Human?

Conversational UI Principles

Have other questions about chatbots?

Contact us to learn more about the pros and cons of chatbots, how to use chatbots as a design research tool, or how to incorporate a human-centred approach to your chatbot design.

Looking For Other Human-Centred Research Methods? 

 6 Street Team Tips For Impactful Public Research

 Using Ethnography in Design Research Can Unearth Problems You Never Knew Existed

 How You Can Prepare For The Future Using Strategic Foresight

 75 Design Thinking Tools and Resources Explained

Previous
Previous

How to Use Community Prioritization Boards to Make Decisions

Next
Next

Are Street Teams the Right Engagement Tool for You?