One of the biggest challenges over the years as new technologies emerge has been deciding what technologies and which vendors to trust. This is particularly true of emerging technologies like artificial intelligence (AI) and one of its most common use-cases, notably chatbots.

Getting Your Technology Right

For their recent annual State of Branding report Bynder surveyed 500 marketing and branding experts on their top challenges, insights and what emerging technologies they plan to invest in in 2019. For the second year in a row, “identifying the right tech to serve as an extension of their brand” was the most frequently cited challenge when it comes to the adoption of new technologies. This was particularly true of technologies like chatbots, voice assistants and augmented reality (AR)/virtual reality (VR).

The research showed that while it is challenging, 94 percent of respondents feel it’s “very” (63 percent) or “somewhat” (31 percent) important for their company to invest in new technologies in 2019, with customer experience and engagement proving to be the number 1 motivation. 

Even though experimental technologies like AI (45.1 percent), VR (36.3 percent) and voice assistants (40.3 percent) are firmly on investment radars, more established outlets like mobile apps (65.7 percent), progressive web apps (50.1 percent) and digital asset management software (46 percent) still make up the largest slice of most marketing budgets. 

While investing in emerging tech is important, it’s proving to be complicated, so marketing professionals aren’t yet all-in on investing in the latest tools available. For now, they’re leaning on more established outlets in an effort to engage with customers and stand out in a competitive landscape. 

Related Article: Top 14 Chatbot Building Platforms of 2017

How Are Chatbots Being Used?

So where are chatbots being used? Robb Hecht, adjunct professor of marketing at Baruch College in New York City, argues that the future of the website is looking like an experience where a user is typing with an AI chatbot who either greets the user as new, or knows a little bit about the customer’s previous history with the brand based on data it has assessed.

As website customers are now in the age of instant gratification, with everyone using mobile phones buying from Amazon and direct-to-consumer subscription startups, customers now expect seamless experiences from brands they interact with on their phones. In fact, research indicates consumers today are becoming more 'experience loyal' vs. 'brand loyal' — as tools like chatbots facilitate the need for speed and response. “Companies are finding that if they don't provide a frictionless experience, consumers can easily find another brand that will. Knowing this, companies are increasingly using AI data backed chatbots to bring us a frictionless user experience,” Hecht said.

Keep in mind there are basic chatbots that offer potential customers generic information such as linking them deep into a website, or offering a customer service email address or 800 number. These types of chatbots, he said, are excellent for brands that need some interactive tool to respond to users evenings and weekends.

Then there are the more significant chatbots that provide deeper services like empathy and motivation. A banking app like Penny for example, provides more than just links and 800 numbers, she asks questions, learns from your responses — and gives you the feeling she has empathy and knows you.

Hecht points out that while users are growing increasingly fearful of how AI technology is overtaking human skills and jobs, they also dearly want AI technology to be human and personalized. Chatbots are customers' first and most direct experiences with AI and demand for them is going to increase dramatically in 2019. Soon, customers will demand that information and personalized experiences be served up to them through AI, as opposed to brands asking them to manually click through content heavy websites. 

Related Articles: How Much Does It Cost To Build an Enterprise Chatbot?

Learning Opportunities

Extracting Data With Chatbots

However, there's more to chatbots than customer experience. One of the best use cases for a chatbot is simple data retrieval. More and more people are looking into diverse data sets, but have less knowledge about where that data might live or how to author the right arcane SQL queries. Zack Nolan, senior vice president of technology programs at Beyond Limits, pointed out that the biggest problem for implementation is solving these two problems in the first place:

  1. Identifying the disparate data - While chatbots can certainly help less technical users interface with and understand patterns in data, the bad news is that most chatbots are not sophisticated enough to actually carry on conversations.
  2. Understanding the common queries - If the expectation of AI is to translate English into SQL queries with some analytics on top, then chatbots might pose a viable entry — no doubt companies are already adopting these systems.

Common user expectations still outstrip the capabilities of chatbots, leading to disengagement or abandonment by users — think of how much more common it is for someone to say they hate Siri than to profess their love for it. “Overcoming these obstacles, potentially by implementing the right user training, will be required to truly bring the chatbot AI into the everyday enterprise. AI and machine learning are enormously powerful tools that can transform many enterprise operations,” he said.

That said, chatbots are not the best UI or UX for sophisticated systems dealing with high-value data and processes. More important than chat is the ability of the AI system to explain itself and its reasoning to the user. Explainability (X-A) is key to building trust with users.

Sabrina Atienza, CEO of Valued, has experience building an AI Slack chatbot to help teams prevent bullying and harassment, and encourage sharing of authentic feedback. She said that before even thinking about deploying chatbots, there are two questions that enterprises need to respond to.

  1. Just like any other software, deploying an AI chatbot will require change management, training and a clear use case. Who is responsible for aligning and communicating all three? Without someone responsible, it's unlikely the AI chatbot will be adopted within an organization.
  2. Is this real AI that learns and improves over time or is this just a scripted chatbot that will require in-house resources to manually configure and update on a regular basis? Most AI chatbots are actually the latter, requiring time, people and effort to add more conversational branches. Keep this in mind when you consider total cost of ownership.

What Enterprise Leaders Must Decide

The problem goes beyond just this. Andy Peart, chief marketing and strategy officer at Artificial Solutions says that enterprise leaders really need to pinpoint precisely where the chatbot will fit into the enterprise and that this should determine what is being deployed.

The need to decide, for example, which platform and channel they are going bet on. Will it be Alexa or Google Home? Will Messenger be the channel of choice? What role will the web, mobile or wearables play? The truth is, your customers want to use them all — at different times and for different reasons — and they want a consistent experience that ideally remembers what had been said in previous conversations. 

The solution is for enterprises to be able to build conversational intelligence once, and then deploy across multiple channels, platforms and languages. When people communicate in a natural, conversational way, they reveal more than just the words they’re saying; their individual preferences, views, opinions, feelings, inclinations and more are all part of the conversation. When many thousands or even millions of conversations are captured and analyzed, this data becomes a unique and powerful source of customer insight, the value of which can be further enhanced when it's cross-referenced with other sources of knowledge.

“Ultimately, if chatbots don’t deliver a true conversational interaction, one that delivers contextual understanding that is consistent across different channels, in any language, it will be a failure for your enterprise,” he said. True conversational interaction allows for a greater customer experience. It also helps to increase brand engagement.