In the early days of chatbots, there was, as with all breakthrough technologies, a lot of hype around how they would transform business as we knew it. Now, two or three years down the road, we’re seeing that these promises were somewhat inflated and premature. The hype has diminished, reality has set in, and chatbots need to start solving real-world business problems or they will be relegated to a business technology fad that failed to find a practical application in the enterprise.

Despite the technology’s potential, it seems prudent that companies avoid a full-scale dive into chatbots without careful consideration. While the promise of AI-enabled assistants and customer support agents is appealing, they are not the panacea we might like them to be.¬†Only the chatbots that help with everyday business productivity and enhance the business-to-consumer experience will fulfill the technology’s promise by “crossing the chasm” (to use Geoffrey Moore’s terminology) between early adopters and mainstream acceptance.

The chatbot as underlying technology

Organizations across all industries are implementing chatbots. Western Union and Deutsche Bank recently rolled out their own chatbots for customer engagement roles, and cybersecurity companies are turning to chatbots to bolster customer support activities.

However, companies would be wise to not treat chatbots as some sort of AI-enabled automation panacea. There’s a reason why many companies that launched as chatbot companies have pivoted to become companies that offer solutions for domain-specific business problems. Companies are beginning to understand that chatbots can be a powerful underlying technology, but that the focus needs to be on solving a specific problem rather than simply investing in chatbots as a catchall solution.

Take, for example, Exceed.ai, which offers an AI assistant that helps sales and marketing teams engage leads. Or Talkpush, which uses “Stanley,” a chatbot, as part of their talent recruitment solution for human resources. Both examples are very specific use cases aimed at addressing singular pain points.

While I’m not the first to say it, I think this approach makes the most sense. These companies — the ones that focus on chatbot products that perform specific tasks exceptionally well and deliver remarkable results — will be the ones to survive in the coming years. Ultimately, the only way chatbots will be able to cross the chasm is for the technology to function as a part of an underlying solution that is focused on addressing specific business problems.

Awaiting the voice revolution

The idea of companies employing a voice interface for chatbots is intriguing. However, we’re not yet to the point of successfully using chatbots for open-ended support and sales discussions via voice. The technology isn’t there, and the industry isn’t there. Right now, what voice-enabled bots are capable of doing is limited. They can deal with simple automation of everyday tasks in exchanges such as:

  • Do you want to change your delivery time? — yes or no.
  • Here’s a reminder of your 4 p.m. appointment in San Francisco. — Press 1 to confirm, 2 to reschedule.
  • Do you need transportation to get there? — yes or no.

Voice bots are not yet able to manage more open-ended interactions. While often-heard statements such as “bots will move to voice because millennials demand voice” and “internet search is moving to voice” may prove to come true, these are not necessarily the proof points that will show that voice bots have crossed the chasm. I believe that the potential for voice-enabled bots goes beyond internet search and other predictable uses.

I think that the most valuable purposes for voice bots will emerge from very specific domain use cases that mandate the voice interface. Scenarios such as field technician communications, workplace collaboration while driving, and emergency alerts are where I expect voice bots to prevail.

Even then voice bots face an additional specific challenge: It is hard to maintain context between distinct voice interactions. Voice technology companies need to facilitate this so the interface is useful for ongoing voice bot conversations.

So where does that leave us? With the industry having entered a “face the reality” phase with chatbots, it’s important that we temper our expectations and accept that only certain chatbot use cases will survive.

Strategic investment is key

According to a recent report by Gartner, by 2021, over half of enterprises will spend more each year on bots and chatbot creation than on traditional mobile app development. If that’s the case, then it is critical to the evolution of this technology that the strategists and developers driving the industry get specific about the real-world business problems their technology will solve in the near future. That is the best way to ensure that chatbots make a positive impact on organizations without companies relying on them to do more than they’re capable of.

Michal Raz is vice president of global partnerships for Nexmo, the Vonage API platform.