The Drum Awards Festival - Extended Deadline

-d -h -min -sec

Artificial Intelligence Chatbots Technology

#SMWNYC: How brands can integrate bots into the overall customer experience


By Lisa Lacy, n/a

March 1, 2017 | 8 min read

Bots are still relatively new, which makes them mostly uncharted territory for brands. That’s according to Dan Gardner, executive creative director of creative agency Code and Theory, who spoke Tuesday (February 28) at Social Media Week in New York.


Ready or not, bots are coming -- and they're only going to get more complex, human-like and conversational.

But because a brand is the sum of all its interactions, brands should not just think about a bot experience, but rather about integrating bots into the overall brand experience, he added.

And while stressing bots are still immature, Gardner said he defines them as artificial people/creatures that hold conversations.

That being said, Microsoft’s office assistant Clippy was ahead of its time, he added. That’s because Clippy was a conversational interface with contextual awareness that could delegate services. Since then, bots have evolved to the likes of Alexa as they develop new ways to interact and become more human.

“They’re going to where you are and becoming ever-present,” Gardner added.

Whereas apps have features in which users perform an action like clicking a button and something happens as a result, bots have skills. That, in turn, makes the decision unique to the circumstance and so developers need to figure out how to provide said skills, Gardner said.

Alexa has 10,000 skills, but they’re still pretty rudimentary, like, “play this,” or, “at this time, do this,” Gardner said. Even Clippy was more advanced. And that means no one is really taking advantage of what bots can do – which is in part because bots are so complex, he added.

There’s a spectrum when it comes to how brands can tap into bots, Gardner said. The core bot skill is rules-based, but, from there, the next layer of skills is conversational and the third layer is proactive and contextual. In other words, in that third layer, the bot will do something because it knows what the user is doing or how the user is feeling. And weather and location-based ideas are the first natural iterations of this, Gardner added.

Finally, the fourth layer is a combination of all of the above in which the bot performs tasks when needed, but also serves as a conversational interface and anticipates a user’s needs at all times.

Media brands in particular have been at the forefront of bot adoption in part because they were forced to embrace advanced technology in creating and distributing content. But, Gardner noted, a well-worn wisdom these days is that brands need to act like publishers and so it would behoove them to look at bots now, too, adding a “bot-first mentality” may become the next digital marketing catchphrase.

“It’s about how to be more human and connect in different ways,” Gardner said. “And how can a bot make my content feel more personal and customized?”

In the past, brands were limited by the technology of what a bot can do, but it will get better in the future. Gardner also cautioned against focusing on perfection with bots that could set high expectations and underdeliver like apps did in early days trying to be everything to everyone.

Instead, imperfect bots are okay early on if they deliver a satisfying experience, Gardner added. And brands that know their bots are imperfect can even play up the personality of the brand in delivering responses as a result, like when a user says thank you to Siri.

“It’s still a master/butler mentality, but that’s the tone,” Gardner said. “Siri knows it is not perfect, but it will offset [imperfection] with a personality that works with the brand.”

Examples include Quartz, which delivers news in what Gardner described as “a more human-like way,” as well as Pogo, which he said provides free, humorous bots that provide “emotional connections to cheer you up.”

“It’s a bot that’s a dynamic therapist,” he added.

It also includes things like weather and a bot like Poncho, which differentiates itself from other weather data providers by integrating personality.

And Code and Theory itself developed a Kanyebot for Social Media Week 2016.

“We’d take the talks that are happening – and created a personality-driven bot to comment [like Kanye West],” Garnder said. “We were tempted to do it for Trump this year, but backed off.”

In other words: Bots can be as simple as entertainment, Gardner said.

But brands must also consider what skills a bot can provide that would improve the customer experience.

“Giving information is one skill, but how do you improve the experience?” Gardner asked.

This requires a shift in thinking to service design in order to add more value as a brand. Further, the quality and value depend on the complexity of the context your brand can provide a solution for, Gardner added.

One example is Domino’s Anyware in which the brand tackled the pain point of ordering.

“[Customers] only think about it when [they’re] hungry, so their strategy was ‘Why don’t we just be pervasive on every platform and go wide and play on accessibility?’” Gardner said.

Sephora’s Kik bot is one that improves the purchase with a product selector that acts as a customer service representative similar to a makeup artist or an expert you could interact with in store, Gardner said.

And the New York Times 4th Down Bot found a way to tell stories quicker and deeper by providing live analysis of every fourth down in the NFL.

“Bots can do something people can’t do, which is interpreting data [like this],” Gardner said. “It judged immediately whether the coach made the right decision. Is this better than an actual coach? It’s telling deeper stories and providing more value that no human can do.”

And then there’s the Stealth Mountain Twitter Bot, which provides proactive assistance and fixes grammar mistakes.

But bots are only going to get more complex, human and conversational.

“As we move through the spectrum, some things will get easier to execute on and some things will get more challenging,” Gardner said. “In the race to reach a person in innovative ways, [there will be less of] the master/butler mentality…and more of a user/bot paradigm.”

Gardner said it’s like social in which brands are encouraged to have conversations with users.

“It’s the same structure within bots,” Gardner said. “Don’t just talk and react, be part of their lives.”

Further, brands can’t stop at just natural language processing. Instead, think about all of the connection points like Alexa and Siri and how to build systems to approach and solve problems as part of ongoing optimization and relationship-building, which takes audience management to a new level, he added.

This, in turn, will mean building out a new set of brand guidelines and character rules specifically for bots.

And, he warned, brands that don’t get ahead with bots will quickly find themselves behind.

“One thing we’ve seen at Code and Theory…is companies that don’t tackle distribution channels early, find themselves in awkward/risky positions in which there is a massive hole and they are too far behind or it takes too long,” he said. “There are so many different reasons why, but if you’re not moving with technology, it will be a problem – especially as it advances.”

Indeed, Gardner said the future is about the combination of bot skills that will continue to make bots more useful and humans get desired responses from bots and vice versa. But the infrastructure needs to be built now – including language processing and distribution.

But, in the end, the result will be better customer data resulting from one-to-one interactions, which will in turn drive brands to become even more personal and contextual, he added.

Artificial Intelligence Chatbots Technology

More from Artificial Intelligence

View all


Industry insights

View all
Add your own content +