Taking a consumer-first approach was a key theme at the World Summit AI, which brought together over 2,000 delegates to discuss applied artificial intelligence (AI). AI technologies are rapidly developing with a lot of applications to already consider and much more disruption to come. Leaders from top universities and companies – including Google, Amazon, Facebook, Intel, Apple and Uber – came together to share existing AI short-comings, current applications and what the future could hold.
The topics covered at the conference were diverse. NASA showed how it is currently deploying AI to push their space exploration boundaries. Swedbank announced that their chatbot will be the first point of contact for consumers across all of their channels by 2020. Pinterest demonstrated how their image recognition capabilities have advanced from simply finding similar products in pictures, to being able to predict objects of intent.
The main advice from the centre stage was to start implementing AI now. It is through the learning and testing that the needed skills and competitive advantages for the future will develop. With that in mind, below are three simple themes from the event that brands can start adopting now.
Understand the business problem first, then add the technology
As Gary Marcus, Professor of Psychology and Neural Science at New York University, shared AI is not nearly as developed as we would like it to be (see image above). However, there is still a wealth of opportunity brands can harness from these emerging technologies.
Describing their AI journey, eBay’s RJ Pittman, SVP, and Japjit Tulsi, Chief Product Officer and VP Engineering, talked about developing a solution that achieved their long-term goal of making their product search easier and more efficient for consumers. Starting small and working in stages, eBay launched the shopbot. At the beginning, the bot had very narrow functionality, using natural language processing to answer simple questions. Over the course of months, the team added on additional capabilities from voice to image recognition, facilitating more complex task and easing the task of finding the perfect product.
There are definitely early adoption advantages to be gained, but it is more important for companies to identify key business problems and future goals. Once there is a clear vision, brands can start adding the different technology solutions to help them achieve their long-term goal in stages.
Work together: the best results are collaborative
Just as working to an objective is important, involving a varied team of experts is also crucial. AI technologies are continuously developing. As a result, brands should be looking to collaborate, not just internally but also outside of the organisation. Working with academics, agencies and consumers will help pose a range of considerations and potentials, as well as foresee and troubleshoot problems.
Martin Kedback, Head of Channel Management at Swedbank, talked about how his company took a slower approach to adopting a digital assistant. Starting in 2015, the brand worked in-house and with third party integrations to build the needed skills and ensure accuracy. What would have taken 30 employees to accomplish is now being handled by three employees with the bot only integrated on the company’s website and app. These efficiencies will only increase as the brand rolls-out their chatbot across all channels, becoming the first point of contact for all consumer engagements. By taking the time and collaborating, Swedbank is creating a solution that both enhances customer service, as well as improves the company’s efficiency by answering simple or repetitive inquiries and ensuring all queries that do need employee intervention are streamlined to the most appropriate person.
Google also shared their view of collaboration, which involves democratising AI through APIs. Lukman Ramsey, Cloud Solutions Architect at Google, illustrated this point by using an example close to home. The company started rolling out their AI capabilities across their own product categories, which saw a huge uplift in new uses and applications in as little time as a year. By opening their APIs to the public, this could further accelerate the adoption and functions of AI.
Take a people-first approach
Throughout the conference, we were reminded that we cannot ignore the end consumer, both internally and externally. Steve Chien, Head of Artificial Intelligence Group, JPL, NASA, knows how hard it can be to get internal backing. His guidance was to know exactly who your internal end-user will be and make the solution relevant to them, as well as focus on solving a challenge that humans aren’t good at.
The Engineering Manager of Visual Search at Pinterest, Dmitry Kislyuk, described how they used image recognition to bring pictures to search. Through the adoption of AI and learning what people were looking for in their picture searches, Pinterest has been able to expand their proficiencies. They are now able to successfully match objects, styles, material and environments. Additionally, they have also developed the ability to identify objects of intent in images, expanding the knowledge they have to increasingly improve their consumer experience.
At OMD EMEA, we have built a process to investigate the possibilities, potentials and pitfalls of AI. From concept to product in two days, we foster a hands-on experience for clients to test and develop consumer-facing AI solutions using the latest technologies. Participants are guided through the initial stages of building an AI activation in a co-creation environment with consumer feedback inspiration throughout.
At this point in time, we don’t understand the full extent to which AI technologies will have on consumer behaviours, but we do have the smartphone as an example. The smartphone drastically changed our behaviour in very profound and subtle ways. We should expect to see the same of AI – just think Facebook newsfeed and autocorrect.