2019’s sure been a year. For South Africa, that means extreme highs and depressing lows, but one things for sure, the country didn’t stop…
2017 is set to be the year of artificial intelligence (AI). With companies such as Amazon Web Services offering AI on tap — including image recognition and natural language understanding — an increasing number of machine learning powered services will be finding their way to the consumer.
Those involved in marketing and retail might have already seen offerings filter through, with the following pointing the way the AI winds are blowing.
Elementary, my dear
In the world of AI, Watson is one of the best-known names out there. Built by IBM, this computer system is able to learn from massive amounts of unstructured data, while also being able to answer natural language questions. Already having made an impact in the healthcare industry, Watson is bringing its AI capabilities to a number of other sectors too.
In the retail space, Macy’s in the States has teamed up with Watson, via intelligent engagement platform Satisfi, to create a mobile tool which allows customers to ask questions relating to participating shops. Utilised in Macy’s On Call, a shopper might ask where to find women’s business clothing, and receive an immediate answer — without human interaction.
Apparel manufacturer Under Armour tagged Watson to help create insights for its online health and fitness community. Through this collaboration they are aiming to provide customised training programmes based on fitness information collected through the Under Armour apps.
Use Amazon’s code for free
With Amazon generating 30% of its revenue through recommendations to shoppers, why would the world’s largest online retailer provide the technology behind this as a free open source project called DSSTNE? Amazon believes it can drive machine learning past the mainstays of image recognition and language understanding, into areas such as product recommendation.
For Google, the amount of data they have available to train their AI to correctly identify a football, for example, is immense. When it comes to product recommendations, this “training data” is far less, and that’s an issue. However, through open sourcing DSSTNE, Amazon hopes the solution might come from outside the company, and they will directly benefit from any improvements to the open source code. And it’s not just Amazon doing this, with Google open sourcing their AI engine TensorFlow, and Microsoft providing their Cognitive Toolkit for free.
The writing’s on the wall
On a more practical level, Dominion Dealer Solutions, a US-based car dealership, turned to Quill, a natural language generation platform, to help write engaging descriptions of their vehicles online.
Powered by artificial intelligence, you won’t find any copywriters lining the hallways at Quill. Narrative Science, the company behind it, believes that the content Quill produces is on such a level that it is “indistinguishable” from that generated by humans. And it seems to be the case, with Quill used by Forbes to generate company earning statements, and Credit Suisse for investment research coverage.
Far from these corporate giants, with Quill writing engaging car descriptions, Dominion saw a 50% lift in page views for used vehicles and a rise of US$1-million in annual revenue. And with similar offerings available, such as Wordsmith from Automated Insights, the days of automatically generated copy are upon us.
New jobs too
While one of the concerns about artificial intelligence is imminent job losses, there will be new opportunities and jobs that flow out of AI. Take as example the ability to combine the capabilities of Quill, and put it to work alongside Molly, an artificial intelligence web designer created by The Grid. Molly creates custom sites crafted around the content and images provided, without the help of humans.
In the near future, it may well be that boutique travel agencies provide customers with unique websites automatically built around the packages they are interested in. For those looking for a trip to Croatia, a custom website could provide information about things to do in Dubrovnik and Zagreb, restaurants close to their accommodation, and activities customised around the specific time they are visiting.
Artificial intelligence doing the media buying
Any marketer stuck trying to build a digital campaign will know it’s a convoluted process. Programmatic media trading has in a way automated the buying and serving of digital ads through an exchange that connects advertisers to publishers.
While this has sped up the process, Wise Data Media believes its AI powered real-time bid management system, dubbed Wise Air, gives them an edge. Able to work its way through consumer data in a way that no human can, Wise Air can help advertisers optimise campaigns, based on getting the correct target market at the right time and through the right medium, all before bids are placed on inventory by the advertiser.
And from here?
While the impact of artificial intelligence is slowing being felt in different types of industries, what is becoming increasingly clear is that this is only the start. As knowledge about AI grows and continues to be implemented in the workplace, we are heading down a path for which the destination right now, whether good or bad, is unknown.
Better buckle up, it’s going to be an interesting ride.