"Every brand will have conversational AI," declares Aaron Shapiro, CEO of Huge, the Brooklyn-based experiential agency for such clients as Google, Nike, and the White House. Shapiro and his colleagues have publicly staked out that the future of digital interaction is the "smart internet," where our experiences are ruled by machine learning automating services and anticipating desires. To help Huge understand where this world of intelligent bots and other AI is going—to aid its clients in being "insightful and not creepy"—the firm has gone so far as to build its own internal AI client. "We're eating our own dog food," says Sophie Kleber, Huge's innovation lead.
Dakota, as Huge's service is known, assists the agency's employees and clients with administrative tasks and finding relevant information quickly. For example, Dakota can pull up information from the company's internal knowledge-management system, such as a presentation for a client. It can identify the correct person or people with expertise in a specific field like retail, and it can help arrange a meeting with all of those people. The meeting rooms in the agency's headquarters are also equipped with temperature sensors which are then connected to Dakota, so if a space is running cold, Dakota can remind attendees to bring a sweatshirt to the meeting. "One hundred million dollars has been spent [in Silicon Valley] just trying to solve 'Let's set up a meeting,'" says Huge's head of business strategy, Bernardo Rodriguez.
A small team of five people spent three months building Dakota, which currently exists as a phone number. Users interact with it via SMS messaging. "We're experimenting with voice and visual identification," says Kleber, who indicates that Dakota will roll out to the whole company in about a month.
The Dakota project has already proven worthwhile for Huge because it shows the potential to increase productivity and improve employee happiness. (Kleber believes that these kind of intra-company AI services have a lot of potential.) More important, though, Dakota has created insights into what consumers really want when we interact with a machine. As Kleber says, "How do we get away from the idea of AI as manipulating us and ultimately trying to kill us to AI as a service? The littlest things go a long way. I say hi to Dakota and she replies to me with the sake emoji. We need to move from Ex Machina to Big Hero 6."