Moving Beyond the Gimmicks with Chatbots - eMarketer
« Return to Mobile Website

Newsletters Sign-Up

Plans & Pricing

Does My Company Subscribe?

Moving Beyond the Gimmicks with Chatbots

November 15, 2017 | Mobile | Marketing

Caroline Klatt
Headliner Labs

When integrated into the chatbot experience, artificial intelligence (AI) has the ability to gather powerful customer data and drive sales, but many marketers are stuck on the idea of a one-time-use bot supporting a big campaign. Caroline Klatt, CEO of chatbot software platform Headliner Labs, spoke with eMarketer’s Tricia Carr about how AI in messaging environments can help marketers in more ways than they might realize.

eMarketer: Using AI in chatbots has been talked about in the industry for some time, but what’s happening now? What advances have been made?

Caroline Klatt: Chatbots were this big announcement, but marketers used them for branding activities that weren’t part of their holistic strategy. They were these gimmicky, campaign-specific experiences that didn’t drive sales, and it soured chatbots for a lot of consumers.

Now, consumers understand the value of chatbots, and brands understand they’re a vehicle for converting sales by sending out optimized messages meant to immediately convert consumers. It’s less about having that shiny object and more about being a part of where consumers spend their day—on their mobile phones chatting with friends and using social media.

eMarketer: When you put AI to work within the chatbot environment, what is it capable of doing for marketers? Is it primarily for personalized communication, or does it go beyond that?

Klatt: AI is a very powerful tool. It definitely enables personalization, but it does much more—you can have personalized conversations that pay out.

In our back end, you can see a lot of data. We use AI to analyze customers’ conversations, and we produce data maps where you can see what customers are talking about. The big trends you otherwise wouldn’t have known are being bubbled up to the top.

You can also segment all of your users. As you see conversations in real time and see the trends coming out of these conversations, users are segmented into different cohorts based on what they’re doing inside the bot. Marketers can then send out very specific messages to groups of users that are perfect for them. All of this is done with machine learning.

The machine also tracks conversions, based on the skew and the user profile, and it will start to optimize the products users see based on what is likely to convert them.

eMarketer: Throughout the industry, have marketers realized the potential of AI in the chatbot setting? Or are they stuck on that campaign-specific use case?

Klatt: It’s very early. I don’t see a lot of marketers taking advantage of it. We still see a lot of those one-time experiences. A lot of marketers say, “We tried a chatbot, but it didn’t work,” but it likely won’t do much for them if it’s a fun, one-time experience that doesn’t lead to meaningful sales.

When you’re in a messenger, you should be messaging—not asking users to try on beauty products, for example. Brands that recognize what the native behavior is and then work into a user’s workflow will realize the benefits.

eMarketer: What new or valuable data points can marketers gather by using AI in chatbots?

Klatt: When you have the ability to parse language inputs from customers, that opens up a treasure trove of data. You can learn about what customers directly ask about, what they’re curious about, how they talk about your brand—which is really valuable for marketers to understand brand perception—and what kind of mobile messaging resonates with customers.

You can also see what users click and where they click and convert. Obviously you can do that in other channels, but you can do it in chat as well.

eMarketer: As chatbots gather all this data, is it being fed into a central database so it can be used across an organization, or is integration still a challenge?

Klatt: Integration is not a challenge technically, because you can reformat data with the API files. The challenge is getting your organizational buy-in, and figuring out the right pieces and the right people to get the integration done. A lot of companies we talk to currently have data all over the place. Centralizing data is a big goal right now across the industry. But chatting with customers is a new and incredibly rich trove of data, and it is possible to integrate it right into a data link.


  • Go beyond the articles:

    eMarketer Products

    You've never experienced research like this.

  • Hear from our clients:

    Customer Stories

    Nearly all Fortune 500 companies rely on us.

  • Want to learn more?

    Contact Us

    Inquire about corporate subscriptions today.