Building Chatbots People Actually Want to Use: Customer Perspectives on Chatbots
Thursday, March 5, 2020
By Sarvnaz Taherian, PhD
The method with which websites and other communication channels are created have in many ways been designed to push away conversations, forcing users to adapt and fit into these rigid structures, languages and technologies. Chatbots provide us with a new framework of designing and interacting with our digital and online experiences. These frameworks flip the model that we’ve accepted as the norm - instead of humans trying to fit around the technology, the technology adapts and responds to us. Seeing as we are so diverse in our experiences, backgrounds, access needs and life stages, having technology that adapts to us makes a whole lot of sense!
However, the technology is still in its infancy. We’ve got a way to go in order to fully understand how we can get the best out of each other. Emerging research on the psychological side of human-computer interaction can shine some light on how real customers perceive Chatbots. In this article, we’ll provide insights into true customer perspectives on Chatbots, as without understanding the real human needs for this novel technology, we’re bound to face a lot of avoidable mistakes and failures.
This one is pretty obvious, but important to point out; customer expectations of Chatbots vary based on previous experiences. Chatbots have suffered from the same evil that all new technologies face - emerging technologies have no baseline and therefore customers are as likely to have had a negative experience as a positive one. This means that we need to put a little bit more time into changing their perspective as the technology evolves.
Level of tech savviness
Digital natives, technology early adopters and tech professionals will be more open to using Chatbots. In saying that, one of the major benefits of a Chatbot is that it resembles instant messaging apps that the majority of customers will be familiar with using. This makes it easier to motivate customers to interact with a Chatbot, but more needs to be done to build a trusting relationship, so people will allow them to perform tasks, assist them or provide advice.
Nature of the task will distort perceptions of Chatbot capability
Our perceptions on what Chatbots can do, and what we are comfortable with them doing changes depending on what tasks your customers require. As a general rule, customers won’t want to use a Chatbot if the task has a high emotional load, has costly risks and/or is urgent. For instance, customers may feel uncomfortable telling a Chatbot to send money to a friend but be completely fine with calling customer service to assist or to do it online themselves.
Effort required to use the Chatbot
The amount of effort it takes to make a problem known and understood by a Chatbot, and the level of typing necessary, impacts on whether a person is likely to use one. Customers feel that it’s easier to use a more familiar channel or interface than a Chatbot if the effort exceeds the benefits, or the task feels too complex for a Chatbot to handle and execute.
Another implication is that if the Chatbot can’t help, the customer has wasted time and energy on a task that then needs to be repeated to a human to fix. People feel that going directly to a human for help will bypass the risk of having to repeat themselves, and avoid added frustration.
Knowing how to speak Chatbot lingo
Knowing how to make a Chatbot understand you can be an art. We’ve all experienced this when asking a Chatbot for help and it tells us that it doesn’t understand, can’t help, provides incorrect information or sends us on an infinite loop (cue: table flip!).
Chatbots that tell customers how to speak to them, what keywords are helpful, what they specialise in, or enabled clickable pre-populated answers are a great first step. Customers don’t want to have to negotiate their needs.
Leaving a trace and providing non-bot confirmations
The lack of accountability and trackability of actions completed with a Chatbot can trigger fear in customers. What happens with their conversations, and where they can access records of their Chatbot interactions?
This is important for building trust and confidence in human-bot interaction, as the customer then feels they have a safety net or hold the Chatbot accountable, should something go wrong. It’s particularly important with newer technologies, as mistakes are bound to happen during the teething period.
The closest we can get to getting Chatbots right, is to involve the customer
We’ve talked about some aspects that can impact people’s perceptions of Chatbots and whether they will use them. But how can businesses cater to these underlying issues and mitigate them faster?
One overarching answer is to apply user/human-centred design methodology or co-design. Both approaches ensure that the customer is placed at the heart of every decision made in the research, design and development process. The added benefit of co-design is that the customers actually facilitate and actively contribute to the research and design process - they are the designers themselves.
My favourite example is from Australia, where the NDIS co-designed the Digital Human, Nadia, to help their clients with disabilities to access their online services. They found that co-design “produces natural contextual human-like embodied conversations: the ebb-and-flow of natural conversational interactions that are impacted by illiteracy, disadvantage, and bureaucratic and technical language” (Centre for Digital Business, 2020). Chatbots and Digital Humans have the immense potential to be incredibly empowering technologies - if we get it right.
At the core of it, we need to understand what’s at stake for the customer and remember to provide value-add. If people are content with the status quo, what extra, unique and valuable things can your customers do with the bot rather than other methods of interacting with your product/service? While you’re at it, make the benefits really clear, otherwise they will fall back on the usual ways of doing things, existing habits.
Look for new opportunities and problems that exist in your business, and explore how a Chatbot could be helpful in this situation, but also how it could be a hindrance. Be honest with yourself and team about the actual value. The adoption of new technology is always difficult, and the last thing you want to do is contribute to the problem of abandonment.
Create design sprints and research experiments to answer some critical questions, for example:
Where’s your Chatbot sweet-spot and what tasks fit there? The sweet spot is where customers have a real-world need for the Chatbots, but also provide a more desirable experience? Think of specific use cases and context to prototype and test around.
Which markets and populations are excluded by your products and services and how can a Chatbot unlock these new customers?
How could we make these experiences more human-like and naturally conversational?
The questions you ask will inevitably vary depending on the product/service/market. Chatbots have a unique and important opportunity to make our online experiences seamless, equitable and accessible. As long as we ask ourselves the important and necessary questions surrounding their use cases and to uncover what people value and desire. If we don’t, then we risk repeating the same mistakes, where information is made available in the context of the organisation and its structures or products, instead of the context of the customer.
Want to build a Chatbot your customers actually want to use? Check out the Chatbots we have created and contact us!