It's hard to go online without landing on a website with chatbot integration, and with the potential of these bots ever increasing we're seeing companies adopt them more and more rapidly. However, this isn't to say that chatbots are all good: have you ever found yourself going round in circles with your query unresolved? We spoke to Botanalytics who combine real time conversations with historical ones to help answer the toughest questions about engaged, churnable and retained conversations, to hear their tips on creating a successful chatbot.
Ilker Koksal, CEO and co-founder of Botanalytics shared some of his key learnings with us:
Because the chatbot craze is still relatively new and experimental, users will likely encounter moments of frustration when chatting with bots. While the potential for error is always present, you can minimize the risk by focusing on a few common problem areas. Anyone interested in chatbot development should be aware of these headache-inducing limitations; you want conversation with your brand’s chatbot to be fun, engaging and simple—not frustrating and error-ridden.
You should always be on the lookout for bottlenecks that cause problems for your users. But if you don’t know where to begin to improve chatbot user experience, start with one of these points.
Keep Your Bot Specialized
Chances are, your bot can’t do it all. Maybe the pros can do silly, extraneous activities like tell jokes between scheduling tasks, but it’s best that you focus on just one specialized purpose for your company’s chatbot. A bot that does one thing excellently is better than a bot that delivers on only half the goals it sets out to achieve.
Ask yourself: how does the bot help the brand? What experience do we want to provide users? To improve chatbot UX, develop a clear and researched use case when mapping out conversations. Begin each conversation with an introduction on the bot’s purpose, so users know what they can get out of the exchange.
Give Your Bot a Personality
Creative, fun and stylized responses make a chatbot more fun to interact with. Adding personality to your bot also provides an opportunity to extend your brand voice in a new way. Design multiple versions of each message so conversations don’t quickly go stale.
Remember that form should follow function: a bot’s personality sets the tone for conversation and should augment the bot’s goals. For example, a bubbly assistant can make clothes shopping more fun, while a laid-back, relaxed chatbot might work better for food delivery.
Constantly Look for Ways to Improve
It might be obvious, but it’s worth saying: chatbot development is a constantly iterative process. You shouldn’t just write out responses and call it a day; you need to frequently check chatbot analytics to see how you can improve chatbot user experience. Look for new queries that arise and conversation paths your users take, and be vigilant of moments that trip up and confuse the bot—or the user.
Also, don’t forget to provide an option to escalate conversations to a human agent. This way, a user can get immediate help if they run into issues or find themselves in a conversation loop.
Don’t Be Afraid to Limit Options
It might seem counter-intuitive at first, but limiting conversation options lessens the probability for error and keeps conversation on-track. By presenting users with a handful of contextual options per conversation step, you also make things more intuitive, taking out the guesswork on how users should respond.
If you need a case study, the 1-800-Flowers bot increased sales by ditching the option to manually set a future delivery date on purchases. This is because the option frequently led to errors, which lost sales. When assessing bottlenecks in your chatbot analytics, think of ways to simplify the exchange with contextual options—or scrap some options completely. Sometimes less is more! A persistent menu, available on the Facebook Messenger platform, helps keep conversations simple too.
This tip is also helpful if your bot has relatively poor natural language processing. Manage user expectations up-front about your bot’s limited understanding of certain phrases. Behind the scenes, you can continue building its language base as you receive more data on how your bot is used.
Integrate Your Bot into the Ecosystem
Improve chatbot UX and make tasks run more smoothly by offering account linking. Let users input account credentials to fulfill tasks faster. Account linking helps to:
A bot that relies on a brand ecosystem or third-party services can help users much faster when it already knows who they are, as well as their past behaviors.
Tickets are limited for this event. Book your place now.
28 June 2017, Amsterdam
The Machine Intelligence Summit: where machine learning meets artificial intelligence. The rise of intelligent machines to make sense of data in the real world. Learn from the industry experts in speech & image recognition, natural language processing and computer vision. Explore how AI will impact transport, manufacturing, healthcare, retail and more.
13 September 2017, London
Leading minds in affective computing, emotional intelligence and machine learning will come together for an evening of networking and keynote presentations around the latest advancements and trends in artificial emotional intelligence. Join a multidisciplinary group of experts from computer science to psychology to cognitive science.
21 September 2017, London
The Deep Learning Summit is the next revolution in artificial intelligence. Explore the impact of image & speech recognition as a disruptive trend in business and industry. How can multiple levels of representation and abstraction help to make sense of data such as images, sound, and text. Hear the latest insights and technology advancements from industry leaders, startups and researchers.