In times of the unstoppable acceleration of the digital hype cycle, it is worth taking a closer look more than ever. And if you look at chatbots, first calm down.
Chatbots are not new; they’re just coming out of the woods
Even before Alexa said her first word (a text that works better spoken than written), even before Amazon, Google, and others existed, Eliza was already there. In 1966, she was the first to master (written) conversation between human and machine in natural language – and not even so badly. She was followed by dialogue systems in telephone hotlines, all sorts of whimsical virtual “interlocutors” for a wide variety of questions, and finally by messenger bots in Facebook, WeChat, Slack, and other platforms, including those of KLM and H&M as showcase bots in the business context.
Amazon’s Alexa, ranked as Voice Interaction (VI), was the latest generation of sophisticated artificial intelligence for “consumer electronics” – it’s fun for the user, but doesn’t necessarily generate sales for companies. Of course, advances in technical and data protection are foreseeable. Nevertheless, we are – at least in the context of “Conversational Commerce”* – still in the experimental stage, where the dictum and phrase “Have solution, looking for suitable problem” really fits! But this also means that others are already over their teething troubles. Anyone who joins a company now and gains experience in this field has a good chance of not being surprised by a sudden technology breakthrough.
* Conversational Commerce describes the digital customer guidance via dialogic interaction, e.g. via SMS, messaging applications, web chat - ideally until the purchase.
Tips and tricks for convincing bots
- Stay human: Simulate “typing times” and “pauses to think” in the bot's response.
- Be pro-active: Avoid uncontrolled activation and context-free interference (unless the user has explicitly enabled push notifications/alerts, for example).
- Generate emotions: Show personality and win sympathy via humor or language style.
- Invest in concept work: well-thought-out is efficiently developed, e.g. with the help of dialogue trees.
- Leverage existing AI APIs: e.g. Dialogflow, Microsoft, Cognitive Services
- Get inspired by existing bots: thereisabotforthat.com
Second, chatbots aren’t exciting until they become action bots
Many of the current, mostly writing-based chat bots are actually small UX rescuers who provide navigation help when websites and/or processes are too complex for the user. This is great and certainly a remarkable automation lever for improving service experiences. But if you want more, Voice Interaction bots in particular should not be seen as alternative web user interfaces, but rather as completely new applications.
As far as users see it, it is most important that bots relieve them of a significant amount of work. This could include a bot that can carry out specific actions independently, instead of only answering questions or moderating a product search. Such bots are called action bots – Alexa’s food order skill is a first Voice Interaction-driven example of this. A lot of imagination went into K.I.T.T. (“Knight Rider”, 1982) or Samantha (“Her”, 2013) or the thought of having your own credit card used by a machine. However, there are promising use cases that work even without having authority over an account – we have a few ideas…
Third – anyone who wants to build the conversation of the future will rely on Voice Interaction
The spoken word is still the least effort for the user, even though a lack of visual support at some point can make mutual understanding difficult. But wherever there are need scenarios that cannot be dealt with on a screen or by finger tap, there is still something to do. So, it’s time to take the first, easy-to-understand steps.
The fact that bots will become a permanent feature of customer-company communication is a very realistic prospect, above and beyond all hype.
You don’t have to be part of every trend just because it offers new technical opportunities. This also applies to chatbots. However, where these artificial conversation partners save users real effort instead of only offering a new interface, it is worth looking at. And try out.
Ask your customers (and gladly us) about risks and side effects.