Innovation expert and THINK17 speaker Tom Goodwin poses strong questions on how and when to best use chatbots. From the hallways of innovation and marketing conferences such as Cannes Lions, CES, SXSW and in the pages of most tech publications, enthusiasm for chatbots is everywhere to be found. However, does reality meet hype and are they really a transformational element? Or is this a fad to wait out? At a time when we talk endlessly about how fast the world is changing and where undramatic headlines don’t garner views, we see little nuance in debates about the future. Any new technology, from VR to 3D printing to the Internet of Things, is judged in a binary fashion. Technologies seem either to be “the future,” or dead. We are impatient too; never considering time as an axis, only thinking about the impact now. Also, timing is everything; laughing at home videophones turned out to be wrong, as did laughing at personal electrical scooters. The Segway doesn’t look so dumb, just early. It’s in this context that chatbots need to be examined. I’m known as a chatbot hater, but let’s at least have a more informed discussion — starting with what they are. What are they? Chatbots exist at the intersection of two transformative movements and technology: where artificial intelligence and instant messaging meet. I’ve long believed instant messaging (IM) to be one of the most powerful marketing platforms the world has ever known. For decades, marketers and businesses have wanted a personal yet omnipresent relationship with consumers. IM conversations can be asynchronous, allowing more time to respond, and yet they are one-on-one, making them supremely personal. They allow both sides to share pictures, locations, contacts, and links. IM apps like WeChat allow people to speak to each other, book taxis, pay money, order food, and rate businesses in one single interface. It’s a personal, mobile, operating system. Artificial intelligence is harder to explain clearly but is more interesting when discussed in terms of meaning and philosophy. AI offers a world of improvement and automation and absolutely has the ability to change everything we know in the world. From cars that route themselves, to market- ing one-on-one at scale, to smart cities that use less power ... the world of AI is a profoundly different way to visualize the future. Chatbots are currently where AI and IM overlap. They are based on the ideals and thinking of instant messaging, powered by AI. But, the truth is they are — currently — not that good. The current state of AI is such that (outside of the powerhouses of Google, Facebook, university laboratories and a handful of others) it’s more likely to be an advanced algorithm. The chatbot interfaces of today seem more like a visualized user flow from a call center menu, with limited menus presented to predetermined questions. Selling, serving or promoting?  For sales, generating leads or converting interest, chatbots may not be the ideal choice. For a brand to spend hundreds of millions in marketing, only to have customers passed to a computer in place of the power and effectiveness of human interaction, seems particularly strange. For customer service, the benefits are more obvious, but I’m still not convinced that the technology is good enough. Perhaps I’m old-fashioned in thinking that reducing churn and increasing metrics like NPS or the chance to listen to customers is too valuable to outsource. Marketers spend millions on customer satisfaction surveys and incentivizing salespeople to cross-sell — why ruin a chance to meaningfully interact? For branded side projects, the opportunities are currently greater. Having avatars talk with you seems to be a successful movement at the moment. I don’t want to chat with Nike or Apple, but a trainer from Nike who can wake me up, keep me on track and shout at me, is more viable. For the moment, however, the advice I’d give is to favor little and later over too much too soon. Luxury goods and services seem especially important to approach cautiously, but areas like insurance, finance, and banking seem so reliant on trust that a cautious approach is needed. The value of trying something first is perhaps considerable, but the cost of failure and loss of goodwill are real. More than anything else, your approach should be dictated by empathy. Robo-answering frequently asked questions Bots helping customers with questions is not a bad idea, but are websites really so terrible for this? Can I access an answer with a single query? If I can install a Nest thermostat with a few steps and no mishaps, then let’s make more things that help people serve themselves. Bots for commerce and service Conversational commerce is all the rage, but is this a better way to shop? Time and simplicity are key: reduce steps, kill complexity and destroy choices. Maybe retail catalogues were not wrong after all. My airline impresses me routinely with the speed of its service on Twitter. Any inquiry at any time results in a pretty fast reply, even if it’s to say “please call us.” A recent situation, where a storm headed for NYC caused me to change my flight routing, resulted in 32 Tweets back and forth, ending in a phone call to confirm changes. It’s easy to be impressed, but what if the same technical thinking to make a chatbot was applied in other ways? The same message could have offered me alternative flight options, with just a short reply needed to confirm my choice. This thinking allows systems, processes, computing, and platforms to come together. One day, companies will be set up to have AI and other advanced technologies built into their core: from personalized special offers to 24/7 customer service, to advertising that flows across devices to build a story. Now is the time for companies to think bigger, more profoundly and to establish these forces as something to build a company around. Above all else, they need to consider these new interactive technologies solely as tools for improving customer experiences. If you’re interested in making sense of Artificial Intelligence, Data Strategy and Everything in Between, please sign up to receive the current issue of THINK Review Magazine: The Data Strategy Issue. And, please join us at the Grand Sheraton at Wild Horse Pass in Chandler, Arizona, May 7 to 10, 2018 for THINK 18, the premier conference to discuss AI, Machine Learning and Data Strategy.