Wasn’t 2018 supposed to be the year of the chatbot? What happened?

In our last post, we talked about the difficulty of balancing the excitement that the limitless potential of digital technology inspires against the practical realities of implementation on a large scale, which led us to ask:

Wasn’t 2018 supposed to be the year of AI and the chatbot? If AI, machine learning, and conversational UI are among the technologies that most over-promised and under-delivered, what happened?

Unfortunately, some technologies have the potential to be more disruptive than others, and it can be difficult to predict where those technologies will take us while a technological transformation is still happening around us. Ten years ago, we never could have guessed at where the smartphone revolution would take us and how fundamentally it would change our lives. But as fast as the pace of change has been, it didn’t happen all at once.

So today we’re going to look at the difficulties of predicting the future of chatbots and AI for business.


One of the most widely circulated factoids that gets circulated in 101 chatbot explainer pieces is: “by 2020, 85% of consumer interactions will be non-human”. It was widely cited in 2018 and has also been frequently cited in 2019 predictions – despite that 2020 is less than 12 months away. But is that really a realistic expectation?

It’s impossible to deny that demand for digital customer service channels continues to grow. In 2018, businesses exchanged an average of 2 billion digital messages with consumers each month. However, that volume is dwarfed by the $1.3 trillion that businesses spend on handling 265 billion phone calls each year. Meanwhile, there are around 300,000 chatbots deployed on Facebook Messenger – nowhere near enough to handle that scale of customer interaction.

Further, it’s estimated that 70% of Facebook Messenger bots are failing to handle simple user requests – showing that chatbots are far from a mature technology for providing frictionless customer service. And, given that 60% of chatbots deployed in 2019 will not include the option to seamlessly transfer to a human, it’s not terribly likely that 2019 will be the year that chatbots take the world by storm – at least not to the degree that they replace 85% of human-based customer interaction.

So what is happening? Where did this figure come from, and why does it keep getting circulated year after year?

“By 2020, customers will manage 85% of their relationship with the enterprise without interacting with a human.”

This quote comes from a prediction made by Gartner. However, it actually comes from a brochure for their Customer 360 Summitin 2011. Since then, this prediction has been recycled almost endlessly, without any attempt to reassess as time passed.

It’s also key to note that in 2017 and 2018, this quote has widely interpreted as referring to chatbots and artificial intelligence. However, the original prediction encompasses all forms of non-human automation, not just chatbots.

Further, those who misapply this quote to promote chatbots and conversational commerce are overlooking the actual history of new digital technologies. Chatbots and AI are only the newest technologies to find themselves the “new hotness”.

Before chatbots, people went crazy for apps – which have not replaced websites, as many predicted they would: The majority of U.S. consumers download zero apps per month, and 80% of consumers will delete an app that doesn’t immediately meet expectations.

Are chatbots really going to replace websites? If apps couldn’t manage it, why would bots?

Over-promised and under-delivered

Chatbots aren’t the first technology that have fallen victim to the hype trap. As Digit founder Eathan Block quipped, “I’m not even sure if we can say ‘chatbots are dead’, because I don’t even know if they were ever alive.”

The simple fact is that for some companies, chatbots are not the correct solution. Currently, most chatbots have been implemented as a form of digital IVR, but if most consumers hate IVR, why are we thoughtlessly duplicating it?

That’s not to say that chatbots are useless! We here at InTheChat are big believers in chatbots as supplements to human agents, for addressing low-complexity high-frequency “nuisance” inquiries that human agents would rather not handle. But if you want to avoid frustrating your customers, chatbots always need to have an easy option to hand-off to a human agent with a complete interaction history, so the customer doesn’t have to repeat themselves. (You can read our previous post here for more buzzword-free, actionable chatbot advice.)


Artificial intelligence is the other technology that causes many problems for analysts looking to make predictions. With a technology that has the power to be as transformational as AI, how do you go about making realistic predictions? This may be why one major analyst firm recently predicted that, owing to ever-evolving threats to cyber security, 75% of retailers are expected to adopt AI-based cyber-defense technologies by 2020.

Such a move would certainly seem to be common sense, but is it likely to happen?

Consider that this same firm predicted, at the beginning of 2018, that by 2019:

  • 50% of retailers would adopt an omnichannel commerce platform, along with the top 30% of retailers investing in “foundational platform technologies that are ‘cloud-based, AI-enabled, and composable'”.
  • 40% of retailers would develop a CX system supported by AI

However, the reality is that in 2018, the majority of retailers struggled to effectively use their own customer and inventory data, let alone implement next-generation technology into inventory systems already plagued with inaccuracies.

According to a survey by HRC, 66% of retailers struggle with inaccuracies in click-and-collect shipping, as these systems often inform shoppers that products are available when there is actually no stock in a given location – causing retailers to implement workarounds, such as showing products to be out-of-stock unless more than one unit is listed for a given location. Further, only 14% of retailers surveyed indicated that they currently used predictive analytics.

So are 75% of retailers really going to implement AI-based cyber security in the next twelve months? It certainly seems unlikely, if inventory continues to be such a big challenge.

More tellingly, IT professionals themselves are skeptical about their organizations’ abilities to act on AI: Only 20% of IT professionals believe their organization has the proper skills, talent, and resources to implement and support AI technology.

Chatbots and AI? Love them or list them?

So what lessons are we supposed to draw? Are chatbots and AI useful technologies for business?

We certainly are big fans of chatbots when used for easy, immediately-practical use cases that relieve human agents of high-frequency, high-nuisance, low-complexity interactions they’d rather not handle – so long as there is always a record of chatbot interaction and the option to easily hand-off to a human agent. Are chatbots going to replace human agents? 

The truth is that the strengths of chatbots are different from the strengths of human agents. A chatbot will never get frustrated about getting asked the same question repeatedly, for example. However, bots won’t have the ability to provide the human touch. Nor are they well suited for resolving complex issues that require both analysis and the ability to be empathetic to a customer’s concerns.

In other words, chatbots are a lot more useful when they are framed as a supplement to human agents, not a replacement for them.

As for artificial intelligence, change might not happen overnight, but it is coming. So while we might take issue with the time scale of the predictions being made, it’s certainly indisputable that: “As we approach 2020, businesses must adopt technology to implement personalization through technologies like AI, predictive analytics, IoT, etc.”

Consumers have little patience for companies that waste their time by limiting how they can communicate, and increasingly are willing to pay a premium for excellent customer experience. Brand loyalty is dead, and if companies want to stay afloat they need to adapt to the realities of digital.

If you have any questions about chatbots and conversational commerce, we’d love to talk – simply click here to connect with one of our digital engagement experts.

You May Also Like

About the Author: Anna Kreider