There has been a huge amount of excitement with the onslaught of chatbots and virtual assistants this decade. Gartner predict that by 2020, 85% of customer interactions will be with AI, rather than humans.
But at the same time, companies like Facebook have scaled back on their chatbots, admitting that the majority were only able to fulfil 30% of requests, seemingly a huge failure.
The problem is clear. Our expectations are far from reality. We see the potential value in a new technology, and then experience a crash when we realize what we can’t achieve. To manage AI successfully, and provide valuable digital self-service, business leaders need to understand what makes this technology challenging, and focus on what we can do to push the industry forward.
Focus on Core Value
Bots fail for many reasons, but one that comes up time and time again is user expectations. As a business leader, you need to think about what you want your bot to be able to achieve, and make sure that this element is flawless. Of course, the bots of the future are sophisticated channels within a harmonious, interconnected experience. Users have increasing expectations of capabilities, and the technology indeed has progressed to justify this desire for a seamless interaction. It is no longer enough to provide limited, targeted, restricted interactions. The bar has been raised, and the customer deserves better.
If you are a weather bot, you don’t need to be able to discuss traffic or music. But your customers will expect you to be able to tell them if they need to grab a raincoat. This becomes even more important when you use conversational UI, as the moment you ask your users to talk to something, they will personify it. This makes it very difficult to change their minds once it’s made up.
88% of users are unlikely to come back to your website after a bad first impression. People judge harshly and are known to be critical, so owning its limitations is integral to a successful chatbot.
If your bot is unable to do what’s being asked, there’s no value in being cute or trying to work around this. The best solution is to be clear about what it can’t handle, and then guide users to what it can do instead.
This is where hybrid experiences can come in very useful. Your bot does not need to be a pure bot solution. Perhaps your solution is meant to help users with flight arrival and departure times, but to book a new journey, they need to open the mobile app, or visit your website. Creating a hybrid journey where users can be guided to the right place for their needs creates a richer experience. Far from taking away from the effectiveness of your chatbot, it actually increases and adds value.
Leverage connecting with human agents too where you need, for example when your bot meets with limitations in functionality, or an issue needs escalation or you have simply found from past trends that customers are fine self-serving for everything else but prefer to complete a specific segment with human guidance. The future of AI is in leveraging the relationship between all the channels on a customer journey, and your business’ ability to make these seamless and complementary. The best solutions will also give you visibility to understand when escalation is preferable, and the ease of allowing you to dictate those situations and trigger this transition when needed without losing the customer or the information along the way.
Effective, not Human
A common misconception is that bots should aim to be as similar to humans as possible. With this (misconceived) goal in mind, chatbots are almost certain to fail. Experts have shown that thought and languages are extremely closely tied, to the point where in order to fully emulate human language we would need to teach our AI to emulate human thought. Perhaps one day we might have mastered what Forbes call cognitive computing, but we’re not there yet. This makes the goal of ‘lifelike’ AI not just unnecessary, but unachievable.
Rather than try to copy human interactions, the biggest wins for AI are leveraging the parts that bots excel at. One example which everyone would agree with is a chatbots ability to be ‘always on’. This is something a human member of staff couldn’t achieve, and gives your business a new way to succeed, a greeting for your customer which never sleeps.
Every medium has its own strengths. For humans, it might be the way in which we think. For bots, it is likely to be something else altogether, from speed or personalization, to its ability to integrate with any knowledgebase or seamlessly connect across channels. By focusing on trying to make bots more human, we are ignoring the capabilities which they have in and of themselves.
While we don’t need to try to make our chatbots more human, we do want to be aware of our customers’ humanity. Just as they will make quick first impressions, they are also impressed by the personal touch, or by your bot going above and beyond.
One way to do this is with anticipatory customer service. Think again about the core value of your bot. With a human interaction, there may be obvious next steps to that conversation which we do without thinking. Let’s say a customer asks ‘Do you have these shoes in a size 6?’ Your bot can say a simple ‘Yes’ or ‘No’ which certainly satisfies the core value. But think about the peripherals around this. It could also say ‘Yes, would you like me to add these to your basket?’ or ‘No, should we email you when they get back in stock?’ Being proactive rather than reactive creates a magic moment where your customer feels supported and listened to, and further helps increase efficiency and satisfaction.
Personalization can also help with this. Look at the information you have about your customer and use it to your advantage. This might be as simple as using their name or location. With the right data, it can also include leveraging account information or past shopping habits.
The Information is Already Yours
Whether it’s proactive questions, personalized touches, hybrid customer journeys or defining your core value, these steps all have something in common. Listening. The main driver for success with AI is going to be understanding what your customers want.
You can use the interactions across all of your channels to get the most accurate view of the needs your visitors are looking to fill. This makes the flow of data between your systems critical for success. Every question which your consumer asks, and each interaction they have, regardless of channel, tells you another element of the single voice of the customer. With this data, you can create bots which hone in on what your customer really wants.
Encouraging your chatbots to imitate humans takes away from how effective and useful they can truly be. Rather than confining your AI to human limitations and borders, a change in this attitude will set them free to do so much more.
When you stop thinking about making your chatbots more human, it opens your business up to think about what their real goal is, and how they can delight your customers with magic moments and added value. Giving your bots a key role on a hybrid consumer journey, promotes a richer customer experience, allowing you to simultaneously learn from each and every interaction.