‘AI’ – artificial intelligence – is the latest tech term getting marketers excited with its potential to streamline, optimise and scale marketing processes.
But it worth remembering also that marketers love new technology and, over the past decade, have been treated to the vast buffet of (useful) new ‘tech’ and, some might suggest shiny new baubles.
So, do marketers really understand what is going on inside the machine and how much of the AI conversation reflects genuine innovation, and how much is… well, marketing.
The concept of an artificial intelligence is nothing new, originating in the sharp intellect of renowned WWII codebreaker Alan Turing who defined the Turing test in 1950 to determine if a computer showed true intelligence. Could a computer fool a human into thinking he was interacting with another human, rather than a machine? If so, Turing proposed, that computer could be considered ‘intelligent’. In 1956, computer scientists rallied around the concept at the Dartmouth Conferences, and the field of AI was born.
Ahead of ourselves
Since then, science fiction has moved faster than computational fact, bringing us the sentient and malicious machines of Terminator, The Matrix and War Machine, whilst developments in artificial intelligence have plodded steadily forward, anchored as they are in the frustrating world of real thing, and constrained by computational power.
The good news first, is that large scale parallel processing is now much faster and more affordable, and this is helped developments in artificial intelligence to rapidly move from towards a more commercial reality.
The most obvious embodiments of AI in everyday life are Apple’s Siri and her counterparts Alexa, Cortana and Google Now. These artificial intelligences use natural language processing and an encyclopaedic knowledge base to ‘help’ consumers navigate the vagaries of everyday life from their smart phones.
These utilise the type of AI that matches Turing’s definition of ‘appearing human’ most closely.
For marketing, the applications of AI are perhaps not as advanced in their ‘humanisation’ as Siri and her colleagues, although some of the ‘chatbot’ tools getting closer.
The marketing applications of AI that are market ready and accessible broadly fall into 5 categories.
1. Content creation You might think that generating unique written content is a particularly difficult challenge for a computer to solve – but not with the help of smart content generation tools like Quill, Hemingway and Wordsmith. These tools provide an array of templates and, using a ‘fill in the blanks’ approach with data elements and keywords, can generate passably well written copy quickly and efficiently. Whilst it’s not award winning stuff, ‘ AI generated content gives brands the ability to generate solid blog, website and email content rapidly, allowing them to maintain currency and ‘visibility’, minus the copywriting bill.
2. Ai enhanced PPC buying Although copywriters may be safe from losing jobs to a journo-bot (for now), media buyers may not be so lucky. Albert a digital-media-buying AI produced by ad tech company Adgorithms is designed to make digital media buying decisions on behalf of brands, eliminating the need for a human buyer. “Albert does all the execution work for you” says company founder Or Shani, “…so you really have time to think about strategy and focus on the creative”. Frank, another PPC optimisation tool looks for the most profitable PPC platforms and buys placement in real time auctions, far more rapidly and efficiently than a human buyer can do.
3. Intelligent content curation Like the recommendations engines behind Spotify and Netflix, marketers can also use the ability of AI via similar clustering algorithms to sort and curate content that fits a customer’s needs, wants and context. Under Armour is one of many companies to engage IBM’s Watson to do this, combining user data with a wide range of third party health and nutrition data to serve up customised insights and training plans in Health Box. More focused on social? Rocco can suggest fresh social content that your brand’s followers are likely to engage with.
4. Chatbots The most commonly recognised commercial application of AI, chatbots have traditionally been used for customer service tasks, but can also be used to help brands deliver personalised website content, product recommendations and shopping guidance. Sephora’s Kik bot uses an established messaging platform to engage with website visitors, learn their preferences and direct them to products and pages that might be most helpful.
The bot even uses a set of product questions to help direct customers to the product line best suited to them – a helpful trait, one imagines, when trying to choose between 50 different lipstick lines:
5. Marketing Automation March platforms have been doing it for years, but with AI, marketers can apply a much greater level of personalisation in the hope of engaging customers and prompting more of them to convert or repurchase. Tools like Boomtrain allow email newsletters to be customised based on a customer’s previous brand interactions, whilst Optimove’s AI creates automated segments, and defines the best content mix for each cohort based on lifestyle, purchase patterns and demography.
One thing marketers must bear in mind, however, is that AI – no matter how sophisticated – performs better with human input and a judicious amount of supervision. As the ill-fated foray of Microsoft’s Tay chatbot into Twitter shows, letting an untrained AI loose can lead to awkward consequences. As with any technology, the boring reality is that a careful implementation, good planning and skilled staff can ensure new tools perform reliably and free up valuable human time as they were intended.