Not a week goes by without some new artificial-intelligence-powered product being marketed to marketers. For marketers wanting to gain an edge on the competition, AI can hold a lot of promise. But not all AI is made equal. If you’re looking to test it you can often find yourself lost amongst a whole heap of tools. Some do prove to be extremely useful. Others…. Not so much. So how can you separate the truly intelligent marketing tools from the useless?
What is Artificial Intelligence?
Most people that I talk to about Artificial Intelligence tend to see it as this revolutionary technology that will either dominate our world and doom us all, or turn it into a Star Wars-like universe where we all get our own personal protocol droid. But there’s much more to Artificial Intelligence other than pure fantasy.
Therefore, the first step in using AI effectively is understanding what it actually consists of. Broadly speaking, it mimics some elements of human intelligence in machines.
It works through machine learning. That’s where an AI model is trained on certain sets of data to learn what’s normal, what’s not normal, what’s expected of it, and what information it needs to output.
Historical data is usually used to train the machine — much like how you’ll show a toddler endless pictures of dogs in order for them to understand that a four-legged, furry creature that barks is, indeed, a dog. Actually, a good example of a machine that’s been trained in this way is Google’s image recognition AI. It’s been trained on the millions of images that Google has access to through the Internet. In a high-profile example, the machine is now able to do something that intrigued internet users from all over the world: identify the difference between a muffin and chihuahua.
Once trained on the right data, the machine can begin predicting results or optimising marketing campaigns. There lies the appeal of Artificial Intelligence for marketing.
AI and marketing, a match made in heaven?
Marketers are well-versed in using data to inform their campaigns, so it’s no surprise then that marketing teams are twice as likely to be using AI compared to other departments.
Increasingly cutthroat as a profession, the average consumer sees up to 10,000 marketing messages each day. It takes a highly targeted, highly personal message to cut through all the noise. And AI promises to deliver just that.
There’s now AI tools to optimise the marketing mix, helping to prioritise marketing channels that will offer the most returns for a business. The machine can analyse customer responses to campaigns in real-time, and then alter content according to responses.
One of the first experiments with this came from the fake coffee brand Bahio, set up by M&C Saatchi to measure the public’s response to an AI-powered billboard. Through a form of AI natural selection, ad copy which held a viewer’s attention for a long time was shown to passers-by over and over again. The same principle was applied to imagery. The end result was a billboard campaign that, in theory, had the best copy and creative for the brand.
Drinks brand Coca-Cola has also been experimenting with Artificial Intelligence for greater personalisation — this time with product offerings. It has invented an AI-powered vending machine that changes a drink formula based on audience and location. A vending machine in a school, for example, will offer different drink selections to one in a gym or a hospital.
AI tools to use in everyday marketing
These examples show off some of the more headline-grabbing uses of Artificial Intelligence by marketers. For everyday marketing, however, there are many useful (and less pricey) tools available.
You can bet there’s an AI tool for every type of marketing.
For those working in content marketing there are tools like Atomic Reach that can help understand what content is resonating with an audience. Concured uses Artificial Intelligence to analyse consumer behaviour towards content, to inform future content marketing.
Then there’s services like Unbabel (wink, wink) that uses state of the art Neural Machine Translation (NMT), Quality Estimation and the efforts of a 55,000 strong community of bilingual editors, to deliver a fast translation with the best possible quality for those looking to translate marketing content for different global markets.
Chatbots are another popular option. These offer a way for marketers to engage with customers and offer support quickly, without having to manually sit by a computer all day or use a contact centre. Travel brand KLM has been using bots to issue travel documents and help its agents stick to its one-hour response promise for any issues. Interestingly, chatbots are used internally as well as externally, to deliver timely information to employees.
There are also recommendation engines that can analyse a customer’s past transactions for any behaviour patterns and then use this to suggest products they might like. Amazon is probably the most famous example of a recommendation engine in action, but many major online retailers use one nowadays.
Finally, there are tools which analyse a business’ current customers and find similar audiences. Emarsys is one provider of this service. This allows marketers to sell at scale. AI can also analyse an audience’s response to a campaign to further refine suggested audiences and improve performance. Harley Davidson used this technique to increase its sales leads by almost 3,000%.
When prioritising which tool to invest in first, consider the ones which will have the greatest impact with the lowest initial outlay. Tools which offer ‘quick wins’ are also a good starting point. It is also worth checking that any tools acquired won’t need too much additional training for the marketing team. Critically, any tool chosen must clearly be in sync with marketing and business goals. Otherwise it won’t be able to show any value to an organisation and investment in it (and possibly other AI tools) might dry up.
AI tools don’t stand alone
Artificial Intelligence is designed to work in partnership with humans. It is there to automate the manual tasks that take up a lot of time, to improve targeting, and to increase marketing effectiveness. A human is still needed to act on any insights from the tools, and to optimise its performance.
The technology isn’t there to replace anyone and equally, tools cannot be left to their own devices; AI tools are here to help you do your job better.
When used correctly, AI tools can make marketing a lot more intelligent.