Nonetheless, some bots have carved out a strong niche for themselves and persevered by playing to their strengths. I’m curious to understand what ideal use cases are for bots, and when they can match or exceed equivalent app or web experiences.
Chatbots can often be thought of as replacements or supplements to existing apps or websites. What are some common gripes around apps or websites?
Another way to think about it is: what advantages does a conversational UI bring to the table? When is natural language a better way of interacting with a computer than a well-designed GUI?
So with that all, where are the areas chatbots can match or exceed GUIs?
Customer service is probably the flagship use case for a chatbot. Specific requests that are context-specific (location, time-based, purchase-history, etc.) are good candidates to be offloaded to a chatbot.
Additionally, because downloading an app or navigating to a bloated website are high friction activities, interacting via text can often be a much more satisfying experience.
1–800-Flowers is a great example of a purchase interaction that doesn’t need an app.
In a perfect world, every app would be wildly intuitive and users would never be confused. Obviously that’s not our world.
Bots can simplify a complicated system by responding directly to a user’s needs. Alternatively, chatbots can act as a user’s designated agent, going out and performing their bidding.
Two great examples of chatbots that navigate tricky systems on the user’s behalf (and also tie together disparate data sources) are DoNotPay, a chatbot for challenging parking tickets, and Truebill, a bot that identifies and cancels unnecessary subscriptions.
Stefan Kojouharov calls this a “superpower”: the ability for bots to be personalized.
Websites and apps are meant to solve problems for groups, but bots can solve problems for individuals.
It feels like this could potentially upend the way we as designers interact with users. Designing truly unique interactions with users will require new methods and tools than what we have now, but could unlock far more engaging experiences than what we have today.
Proactive notifications could be an opportunity for bots. Being able to reach out and let a user know something in a personalized or context aware fashion gives bots a leg up on more naive app push notifications, and being able to message a user where they already are (for instance, on Facebook messenger or via SMS) and drop right into the interface is an opportunity to reduce friction.
News bots, like CNN’s bot, are great examples of chatbots that proactively keep the user abreast of developments.
The flipside of this is that when bots inevitably overstep the bounds of good taste, they’ll be treated as spam and banned. Teach your chatbots to be good internet citizens and don’t over-notify.
Before the internet, contacting a business meant waiting until business hours. Once the internet came along, you could interact with a digital storefront at all hours of the day. Bots promise to do the same thing to real time communication with businesses, by making staff available 24/7.
“On a conversation you expect immediate response. Major advances in real time data availability are making the conversation possible.” — Pietro Casella
With little visual design and, increasingly, no need to write a line of code, bots could be an ideal way to test ideas and assumptions on the fly. No need to resort to visual mockups, wireframes, or websites, just launch a bot and collect your insights!
Group interactions, especially among strangers or colleagues, acan be staid. A bot could be a welcome participant for introducing some levity or breaking the ice.
A great example of this from my personal experience is the Giphy chatbot in Slack, ubiquitous in most groups I frequent.
“They hold out the promise of being able to generate more conversation than might have otherwise occurred between the humans. To act as conversation starters/primers. And/or conversational maintainers.” — Natasha Lomas
While undoubtedly overhyped, chatbots still have a number of use cases where they win out over traditional apps or websites. And it feels like we have yet to see the killer bots (no pun intended) emerge on the platform.