More and more websites and apps are integrating chat functionalities. These are often connected to a service agent or a scripted bot to resolve simple issues. However, the traditional and confined chatbox is often disconnected from the application, service, or website. We use it to complain, to ask simple questions, and sometimes to work out more complex issues. Users usually resort to chat because information or features are unclear or missing.

This approach may work in many cases, but it’s worth thinking about improving the traditional conversational experience, making it more meaningful, perhaps even pleasurable.

The Conventional Conversational Experience

Let’s start with a simple example: an e-commerce website selling fashion items. The company offers a refund policy of 14 days, however, there are several rules around this policy. For example, when the user returns a piece, the item has to arrive within 14 days. In other words, sending the clothing back on the 14th day may be too late and could result in not receiving a refund. Additionally, extra costs could be involved, e.g. mailing costs not being covered when the return didn’t hit the timeframe.

Even though you can find the refund policy rules in a long list of terms and conditions, but the user simply reads “14 days return on all items”, buys a dress, wears it once at a party, and sends it back one day too late, in a wrinkled state. Basically, turning the dress into a second-hand piece of clothing. A fairly common scenario for many e-commerce websites with. many buyers being unaware of how exactly the policy works.

After the hearing she won’t be receiving a refund for the item and mail, the dissatisfied customer resorts to the chat feature as it’s the most prominent and fastest way of filing her complaint, hoping to reach a compromise with the agent.

The Pro-active Conversational Experience

What happened in the previous scenario could easily be avoided, for example through a pro-active chat experience. Instead of waiting for the customer to state his or her intent, the system actively offers help and perhaps even warns the user.

Rather than simply mentioning the 14-day return policy during the check-out process or on the shopping cart screen, the user could be notified with a list of simple requirements related to the return policy.

Let’s not forget that we are designing a conversational experience. In other words, we could simply add a list in the content to solve this problem. However, we want to engage the user and capture his or her concerns and problems before it’s too late. This is one of the strengths of conversational UX, allowing the user to “talk” to your digital service. Whether you’re using agent-controlled, scripted, or AI-powered chat experiences.

We have to be careful about when to be pro-active as we shouldn’t be throwing messages at the user on every screen. Instead, consider problem cases, ideally user-specific ones, with the information that we have, and see if we can overcome these barriers through conversational design.

When we think about the UX Pyramid, we’re simply upgrading the conversational experience from usable to convenient. The user should now be well-informed and chances are lower that the user will be in for a surprise when returning the item incorrectly.

From Conversation to Assistant

Basically, the ideal state of a conversational experience is one where we break out of the typical chat window. In concrete terms, the user could switch to a complete “conversational state” of the system.

In this case, we’ll take a more complex example and use an insurance company where many terms and policies can be hard to understand. They are often explained in great detail, but the wording may not always be clear to the user. Furthermore, buying an insurance policy is a high-involvement purchase.

We still use tools to help the customer choose the right product, such as dynamic graphs. However, we’re adding an entirely new conversational layer on top of the experience. The user can still ask questions, for example, to gain a better understanding of a particular benefit or how a policy will impact his or her taxes. But what sets this model apart is having the ability to change and expand content and features to make the experience more consumable for the user.

In effect, the user is controlling the experience, even building the website or application based on his or her needs and level of understanding. In other words, the interface would modify based on the user’s conversational interactions and could potentially end up looking completely different than its default state.

Conclusion

This is a big challenge in terms of technology. However, we’re looking at how conversational experiences could evolve. Will we keep using a chatbox that floats over the website or app? Or will we allow the chatbox to connect and interact with the website or application?

The insurance example was fairly extreme in terms of functionality. That said, small interactions in a conversational environment, such as highlighting content or links, could drastically improve the user experience as the chat functionality is no longer disconnected from the primary interface.

In short, a text-based conversational interface should eventually break through the walls of its chat window and interact with the website or application.