Just this week Stephen Hawking, warned that Artificial Intelligence could destroy civilisation, and yet it is regularly being touted as a useful technology as far as marketing is concerned. Indeed, it’s possibly the most hyped development within the marketing community at the moment. So what’s going on? What exactly is it? And can it improve B2B marketing?
In common with other new technology terms that become fashionable, artificial intelligence is talked about in many contexts, often applied to products and solutions that display very little intelligence, artificial or otherwise. Nearly half of marketers say artificial intelligence is an overhyped idea/term, according to recent research from Resulticks. And given the very casual way the term has been used, an outsider could be forgiven for struggling to understand the term or what exactly it is. Confusion also arises as the term can be used to describe two distinctly different things:
In fact, many of the processes, solutions and software that are sold under the AI banner do just use rules known as algorithms. Algorithms are sets of often complex rules that computers follow when processing data. Let’s take crossing the road as an example, you could build a robot that followed some simple rules to cross the road:
The robot is just following rules – no creative intelligence is required. You could add as many extra rules as you like, perhaps taking into account the speed of traffic, the width of the road, the height of the kerb, whether anyone else is crossing, the best place to cross and so on – but the robot would still just be following rules.
As computers have become more powerful and the cost of storage reduced, these rules can get very complex and can be processed very quickly. This allows for almost magical abilities we’ve not see before – like facial recognition, or self driving cars. But these still use algorithms – rules that humans have created.
And these systems are far from perfect when they attempt to do too much, are pushed beyond what they were designed for or when, all too commonly, their human written rules are incorrectly written, sequenced or otherwise flawed.
Does all this mean they are worthless? Far from it, complex algorithms are incredibly useful and a modern computer system’s ability to deal with truly huge amount of data means that they are ideal for dealing with the deluge of digital data modern marketers have access to. Even without intelligence they can help with a wide range of marketing tasks.
At it’s basic level marketing automation is one the least “intelligent” of technology-led marketing solutions, however it could be argued that it is the most useful. By automating many of the marketer’s daily tasks marketing automation saves time and money. Unlike some fancier artificial intelligence technologies, marketing automation works reliably and is in use across all industries.
HubSpot is perhaps one of the most well-known marketing vendors with an establish and comprehensive product that allows for all aspects of your marketing to be monitored, managed and automated. Marketing automation helps you work more efficiently and improves the engagement and interactions with your clients and prospects.
Email marketing is an obvious area where the size of the data pool and wealth of statistics mean that computer systems can optimise performance. Computers are now being used to augment the traditional A/B testing methods to allow for a wider range of tests to be more accurately rated and systems like Optimail even perform live testing – optimising on the fly.
Other platforms out there look at optimising subject lines, body copy and calls to action; according to one supplier, Phrasee its AI-generated subject lines outperform human-written subject lines more than 95% of the time, and another, Persado, goes so far as to claim that its “cognitive content” outperforms man-made content 100% of the time.
Email platform vendors are now working on systems that learn when each individual subscriber is most likely to open their emails so that each and every email can be individually sent at the most optimal time.
One area that many see as incredibly important is the use of predictive marketing – using big data from your analytics and elsewhere to determine the optimum marketing strategies and actions to follow. According to MarTech expert Doug Karr, 68% of survey respondents claimed they believe it will be a critical piece of the marketing stack, and 82% of companies are researching its use.
To work well predictive marketing needs three elements – a clean set of ‘good’ data, a software system that can do something useful with the data, and a data analyst or experienced marketer to recognise trends in the results and make sensible tweaks and decisions based on the software system’s output.
Here are just some areas where predictive marketing can help:
There are serious concerns over the use of artificial intelligence technologies and many members of the public are uneasy about their use. Take advertising as an example, Campaign has reported that 92% of UK consumers believe there should be regulation with a legally-binding code of conduct, while three-quarters (75%) think brands should need explicit consent before using AI in their marketing. In customer service and support too, the report found that 28% of respondents (and 33% of women) would feel more negatively towards their favourite brand if they discovered it was using AI instead of humans.
Given the sheer amount of personal data out there and the possibility of using computers to link it all together, privacy campaigners are particularly alarmed. So always consider how the customer might view your use of their data.
However useful you find all these undeniably helpful technologies, be aware that they are prone to error and making potentially embarrassing poor decisions. Always keep an eye on what they are doing and find some way of independently checking their performance, accuracy and usefulness.
Even a fairly simple system like the self checkout tills in a supermarket can’t get it right most of the time. “Unexpected item in the baggage area” and “Please wait for assistance” being messages that are all too common. And the FBI’s state of the art facial recognition software – uses an algorithm which misidentifies faces almost 15 per cent of the time.
So whilst many see artificial intelligence as an over-hyped buzzword, the principles and science related to the field have many practical and powerful applications in the world of marketing. Marketing automation, email optimisation and predictive marketing are just three key areas B2B marketers should be investigating and using.
Thankfully we are still quite a way from computers doing ALL our marketing for us – in the immortal words of HAL from 2001 A Space Odyssey: "I'm sorry, Dave, I'm afraid I can't do that."