It’s not surprising that artificial intelligence experienced a backlash as organizations experimented with a brand-new technology. Like wearables, virtual reality, and cryptocurrencies before them, we rushed into chatbots as “the next big thing” — even at the cost of user satisfaction. The resulting cooling of attitudes in many ways signaled the chatbot’s coming of age.

Early bots were fundamentally tools in their infancy that were rushed to production on the crest of the 2017 hype wave.

“Once the current euphoria on bots dies down, we will see the emergence of real solutions to real problems coming out of the chat ecosystem” Bursting the chatbot bubble – TechCrunch

The chatbot excitement-bubble is well described by the Gartner Hypecycle. Throughout 2016 and into 2017, chatbots were riding high on the “peak of inflated expectations”. Today we are trudging through the trough of disillusionment — which is probably the most exciting time to be a part of a technology.

Chatbot Hype Cycle

Hype Cycle Research Methodology | Gartner Inc.

As weak implementations abandon ship and early adopters build a more mature understanding of the technology, we become better able to leverage bots to solve real problems. That opens up space to make this technology really work for organizations and their customers.

The path forward

The core value proposition for a bot is two-fold: they simplify and expedite the user experience while automating cumbersome or repetitive tasks usually done by humans. It is critical to choose the right tool for the job. When used in situations where they don’t add true value, chatbots lead to user fatigue and frustration.

When users opt-in to a new type of experience that requires them to engage in uncomfortable behaviors and provide a lot of personal information, their expectations for a good payoff are pretty high. Users expect personalized, human-like interactions with bots — and that’s frequently where they fall short. Chatbots need to not just understand meaning, they need to learn from their mistakes, understand context, and analyze user sentiment. There are four major ideas that we like to focus on:

  1. Don’t try to be everything to everyone. While artificial intelligence and natural language processing are getting better by leaps and bounds, chatbots still excel when focused on a handful of specific, meaningful tasks.
  2. Keep a human close by. Recognizing when a handoff to a human is necessary and doing so promptly helps users feel more comfortable interacting with the bot and less frustrated when something goes wrong.
  3. Focus on pain points. A high degree of polish and attention to edge-cases helps prevent catastrophic headlines from frustrated users. Talking to the bot should feel like a real conversation.
  4. Find a true need. Resist the urge to implement a solution without fully understanding its value proposition. Chatbots are a powerful tool that should be used to solve a specific set of problems.