Read our new manifesto

Start 2021 with fresh ideas
and practical tips on...










Banner BGBanner BG

‘We need to be aware as an industry that we could quite easily spoil this for everybody’: why brands shouldn’t jump the AI gun

Taco Bell's TacoBot

The concept of “conversational commerce” has been top of mind for many marketers this year as brands try to figure out how to leverage popular platforms like Facebook Messenger and Amazon Echo to connect with consumers in a way that feels personalized.

Using artificial intelligence (AI), brands including Sephora, American Express and Taco Bell have created chat bots that interact with consumers on messaging platforms like Kik and Slack, while other brands have used the technology to talk to people via voice-controlled devices. For example, people can ask Patron for drink recommendations and recipes on Amazon’s Alexa, while Fitbit users can get updates on their stats and progress by asking Alexa a few simple questions.

At a panel at Ad:Tech New York, ZenithOptimedia’s EVP-head of innovation Tom Goodwin and R/GA’s executive director Cindy Pound discussed how brands are using conversational commerce in their marketing efforts today and why marketers should take caution when it comes to touting their latest experiments with AI technology.

The rise of AI and conversational commerce

The opportunities that artificial intelligence holds for marketers are endless, but many brands are just now starting to experiment in the space. Since it’s still early days, Goodwin said that advertisers should be wary of slapping the words “artificial intelligence” on their products or services at this stage of the game since it could end up confusing consumers who might not fully understand the potential that’s in store for AI.

He explained that while gadgets like Amazon Echo and Apple’s Siri are already proving to be useful for consumers and have been utilized by brands as well, they still have a long way to go when it comes to their AI and machine learning capabilities.

“Pretty much 99.9% of things that proclaim to be AI are not in any kind of meaningful way AI,” Goodwin said. “I’ve told my Amazon Echo every single night for the last year and a half that I’ve got to wake up at seven o’clock in the morning. I say, ‘can you wake me up at seven?’ and it still, every time, asks, ‘seven a.m. or seven p.m.?’ That is not machine learning.”

While Goodwin pointed to IBM’s “Watson Ads” offering - which lets brands give personalized and contextualized information to consumers based upon interactions - as a strong example of AI’s potential, he said that overall the industry is still about two to three years away from really seeing how AI can truly transform a consumer’s interactions with a brand.

“We need to be aware as an industry that we could quite easily spoil this for everybody by claiming that people’s first experience with this stuff is AI,” he said.

Current limitations

Cindy Pound, executive director of R/GA, cited Starbucks’ recent pumpkin spice latte chat bot on Facebook Messenger as an example of a brand that might be spoiling the concept of AI for today’s consumers.

Pound said that while the chat bot ended up garnering an impressive half a million messages, with users spending an average of about two minutes talking to the bot, the coffee chain failed to provide pumpkin spice latte lovers with a truly interactive and fun experience since there were only a few questions that the bot could actually answer in a coherent manner.

“The interactions were limited,” she said. “I think the marketers were thinking [they were] just providing a fun utility, but then there was this consumer-level disappointment because the technology wasn’t as smart and pithy and fun as it would [have been] interacting with someone with human intelligence.”

While these kinds of experiments can provide marketers with a flashy press release and potentially good metrics, Pound said that brands ultimately set consumers up for disappointment when they try to bill these types of artificially intelligent experiences as comparable to ones that involve actual human interactions.

Future of AI

As is the case with many of today’s buzzy technologies and platforms, Goodwin said that the brands that are most likely to succeed with AI down the line are the ones that provide clear functionality for consumers. For example, he said he’s not interested in having a conversation with his “favorite W Hotels character” – instead, he wants the hotel chain to be able to do things like have his favorite channels appear on the TV when he walks into his room.

“What we really need to do with all of this new technology, especially with something like AI, is really embed it in the very fabric and core of a company,” said Goodwin. “Real transformative innovation will happen when these incredibly profound technologies are taken to the very heart of businesses rather than the exterior.”

Pound echoed this sentiment, and added that while things like Twitter’s recent implementation of chat bots into direct messages will help brands make strides, there is still a long way to go.

“Chat bot is a really hot word in 2016 and you see some experimentation and some advances in the technology making it more accessible to develop these programs,” said Pound. “But I think we’re kind of a ways off from them happening at scale or generating long term business value.”

Join us, it's free.

Become a member to get access to:

  • Exclusive Content
  • Daily and specialised newsletters
  • Research and analysis