Mel is having a busy day. Mel is on the marketing team of a large pharmaceutical company and she’s already analysed data from over 1,000 unique journeys on an online patient support programme, written and published 10 approved and compliant social media posts for various platforms, placed online advertising for a disease awareness campaign and made technical adjustments to a company website to ensure it’s optimised for search visibility.
There’s nothing particularly unusual about any of that, except for the fact that Mel has done all of this in the last six and a half minutes and hasn’t even had a coffee this morning. That’s because Mel is a machine. Not the kind of hard-working, driven and career focused person you might figuratively describe as a machine – Mel is an actual machine, made up of a complex combination of machine learning algorithms that could also be described as artificial intelligence.
While this may seem like a far-off fantasy environment, some studies suggest that nearly half of all jobs could be automated within the next 20 years – and that figure includes a number of highly skilled and often well paid consultancy and marketing roles.
Artificial intelligence is starting to become commonplace in many industries and marketers have already started taking advantage of the benefits that high-quality automation and machine learning can bring. Google has been using an AI system called RankBrain to interpret complex search queries which require natural language processing and a deep understanding of context since as early as 2015. A number of companies have also started using AI to prevent security breaches of marketing data by tracking and analysing vast data sets to identify suspicious activity.
Although some of the early forays into chatbots and AI in social media haven’t been regarded as the most successful marketing exercises, AI and machine learning do offer significant potential benefits to online platforms. Companies like Facebook are investing huge amounts in AI research, not just to serve up the best content or most relevant advertising using the huge volume of data they have, but also to interpret and describe images so that – for example – visually impaired people can use its services.
Naturally the tech industry, with its fast moving culture, huge budgets and foundations in technical innovation will lead the way with AI and machine learning – but pharma, too, has the potential capital and innovation-focussed environment to spark massive change. The pharma and, more broadly, healthcare industries are already utilising early AI experiments in a number of fields, from drug discovery to the detection of certain types of symptoms. In an old world where doctors would need to read up to 160 hours of new medical research produced every week just to keep up, computers can scan, index and understand vast amounts of content in a matter of seconds.
IBM has been working with healthcare companies in the US with the aim of using its AI technology, Watson, to improve diagnosis and reduce costs at the same time. Watson has absorbed more than 600,000 pieces of medical evidence, more than two million pages from a wide range of medical journals and has the ability to search through over a million patient records, giving it an obvious advantage over a human doctor – in certain respects, of course.
The next big step is likely to be in marketing because the applications are practically endless and the potential gains so large. The pharma industry offers exactly the kind of complex, large volume data which allows deep learning to offer huge value. Machines and algorithms also follow rules exceptionally well – something of huge importance in such a detail focused and highly regulated industry. In a few decades time, an AI marketer might undertake the types of large-scale tasks in a morning that would be impossible for a human to accomplish in an entire career.
Marketers shouldn’t feel like AI and machine learning are out of reach in 2017, however. Much of the recent work completed in these areas is open-source or available online, even from big companies like IBM. At Pegasus we have a team already looking at how AI can help solve some of our clients’ challenges – for example, building prototypes for complex data analysis, or chatbots to interact with people and engage in dialogue based on its understanding of a user’s personality and emotion.
The potential ethical and socioeconomic concerns presented by the inevitable growth of AI is the topic of numerous further articles. Here, I’d simply encourage anyone with an interest in the subject to begin playing around with a few of the examples already available online. None of us could ever be quite as efficient as Mel, so it’s vital we all start to understand how the power of this fast-developing technology might be harnessed to positively benefit ourselves – and the future of our industry.