Ad agencies are rushing out artificial intelligence services

In the context of advertising and marketing, AI theoretically means more personalized and interactive consumer experience, including targeted programmatic ad buys, identification of site visitors’ decision-making patterns, conversational commerce like bots, as well as smarter search and recommendation engines on websites, according to six agency executives interviewed for this article.

At the moment, with the help of AI developed by big tech companies, agencies are able to serve cognitive ads and integrate voice-activated assistants in their campaigns. For instance, in September, GroupM’s MEC worked with IBM in a cognitive campaign for Campbell’s Soup, where Watson created personalized recipes as display ads for the brand on The Weather Company’s website based on a user’s location, what the weather is like in the area and the user’s ingredient preference.

In the same month, agency CP+B helped Domino’s release a Facebook Messenger ordering bot, and back in February, the agency added pizza-ordering support from Domino’s to Amazon Alexa, where users could ask the virtual assistant to place an order and track their order.

“The most applicable things in AI right now are chatbot and natural-language processing,” said Joe Corr, executive creative technology director for CP+B’s Boulder office. “Utility chatbot is an easy first step. We are looking to experiment with chatbot that can reflect a brand’s voice and offer entertaining experience.”

Of course, another important pillar of AI is creative. Saatchi & Saatchi’s Team One brought a short film that was created, concepted, directed and edited by machines to Cannes back in June of this year. While this may look like a PR stunt, it just shows that agencies are looking to learn more about the language of AI and inject creativity into AI implementation.

“At some point, AI could even create a better ad than a creative director,” said Meeker.

Sure, why not? The thing about AI is that, like “big data,” it is affixed to lots of things. The AI market is estimated to reach $5.1 billion worldwide by 2020, up 54 percent from last year. And the media and advertising sector is expected to take the largest share of the overall AI market, according to research firm Markets and Markets. Agencies see that kind of money and smell opportunity, even though much of the AI market right now is not squarely in the services most agencies provide.

MDC Media Partners officially debuted an agency called Born on Wednesday that is focused on AI. Partner of MDC’s investment arm KBS Ventures, Mike Nicholas found that the problem with agencies is that they are not well-versed in AI but they understand brands, while startups have intensive AI knowledge but they don’t know the commercial applications of their technology. So Born is created to examine every existing framework from different technology companies and stay on top of all emerging AI platforms from the startups. Then the agency can negotiate the ins and outs of the technology based on a brand’s needs.

“If we cannot bridge the gap between agencies’ media knowledge and startups’ technical understanding of AI, we cannot succeed,” said Nicholas.

As Nicholas described, the design process in AI is based on characters and user experiences rather than on graphics or visuals. And creative character development is focused on vocabulary, diction and syntax. For instance, a creative may write copy for a brand in response to everything someone is going to say. The voice has to be reflective of the brand, not robotic. This type of character development is more like writing for a video game development or movie script development, and less like a copy treatment for an ad, he explained.

“We are an agency that literally delivers ‘brand-as-service’ instead of brand as message,” he said.

Born only has four full-time employees at the moment, but it will hire more staff as needed, according to Nicholas.

Huge, on the other hand, developed a virtual assistant called Dakota around two months ago as a use case, letting employees have conversations with AI. Dakota exists as a phone number, and it already started helping manage administrative tasks at the agency’s headquarters in Brooklyn.

For instance, a person can ask Dakota, “Who is the expert on retail?” Dakota may answer “Emily.” If the person wants to set up a meeting with Emily, Dakota can pull her schedule and further book a meeting.

“We’ve learned from Dakota that it’s important to narrow the context in conversation with a chatbot,” said Bernardo Rodriguez, the agency’s managing director for business strategy. “So a brand may not be able to manage general complaints with a single bot, but it can develop a bot to specifically handle billing complaints, for instance.”

MDC and Huge aside, Team One assembled an in-house team of six to focus on AI, where they work with clients and partners like IBM Waston and Google, and address AI at industry conferences. This framework is reminiscent of the agency’s virtual reality lab that was developed back in April of this year. The lab model proves efficient for Team One as it allows the agency to prototype, pull partners and in-house talent who are passionate about a specific topic, according to W. Joe DeMiero, management director of digital for Team One.

“Our VR lab started by two people in the closet, and now it’s a 500-person team,” he said. “I think AI will be the same.”

Of course, not every agency plans to set up a separate entity or an in-house team to implement AI. GroupM and Isobar have run a few AI campaigns for brands like Google and Campbell’s Soup, but the technology doesn’t affect any internal structural changes at the two agencies.

“We fundamentally believe the tactics that AI brings to the table are already integrated into the work we do,” said Cary Tilds, chief innovation officer for GroupM.  “All of our teams must understand AI. We don’t need to form a separate group to show that we are invested in it.”