The overhype of artificial intelligence seems to have died down, replaced with a new clarity regarding its real potential.
AI, more often machine learning, is now being widely adopted rather than merely talked about.
So, how exactly is it being used?
There’s much more on the topic in Econsultancy’s AI and ML Best Practice Guide, but in the meantime, here are a few ways marketers are harnessing the power of the technology.
Data analysis and speed
According to Econsultancy’s 2018 Digital Trends Survey, 49% of respondents say they are using AI for the analysis of data – more so than for email marketing and on-site personalisation.
There’s good reason why. In the research and planning stages, artificial intelligence can automate and improve tricky and time-consuming tasks. For example, instead of a marketer manually sifting through customer reviews and surveys, an AI can instantly process and review the data, as well as provide actionable insights on how to improve performance.
The speeding up of processes is an undeniable advantage of AI, and as a result, we’ve started to see companies use the technology to disrupt legacy systems and provide winning customer experiences.
This is becoming prevalent within the insurance sector in particular. Property and casualty insurance claims have historically taken weeks or months, but US-based company Lemonade aims to revolutionise this by automating the process. In 2016, its AI set a world record by settling a claim in just three seconds.
Picture this: an #insurance policy you can actually understand, in #PlainEnglish Massively shorter read time, fewer exceptions, zero page jumps. Contribute to @Lemonade_Inc Policy 2.0 yourself: https://t.co/Cpvgjpcj0z#GoLemonade #transparency #InsuranceWillNeverBeTheSame pic.twitter.com/kpB6UpJk2o
— Lemonade (@Lemonade_Inc) May 17, 2018
Meanwhile, industries with extensive research requirements are also seeing huge benefits from machine learning.
In the world of law, for example, ROSS – an AI-powered legal assistant – helps lawyers deal with the large amounts of documentation and legislation associated with cases. It continually learns over time, tracking developments in the law and notifying users with real-time updates.
From data analysis comes relevant experiences for customers.
Sport England is a good example of a brand that has used machine learning to change consumer behaviour, using Spotify’s ML-led data to drive its the ‘This Girl Can’ campaign.
It found that women listened to specific playlists when exercising, meaning that when they stopped listening, the likelihood is that they have also stopped exercising. Spotify data was used to identify women who had stopped listening to workout playlists over the previous 30 days or more, with these women then being served ‘This Girl Can’ content to encourage them to start again.
Reports suggest that nearly half of all women took action as a result of seeing the campaign, and engagement was five times the usual Spotify benchmark.
Creativity and content at scale
There’s been much ado about artificial intelligence ‘stealing marketer’s jobs’. This feels more like scaremongering than the reality, with AI actually aiding and improving marketers’ performance in most cases.
One area which both marketers and consumers have expressed concern over is creativity, and the question of whether or not AI will hamper creativity within advertising and content marketing.
Interestingly, however, instead of negating creativity, AI and ML may reassert its importance. Take email subject lines, for example, where repetition (and opinion from others) can lead to a lack of creativity and more time wasting. Companies like Phrasee, which uses deep learning to generate marketing language, aim to take away this kind of human bias.
As well as creating the right message, another benefit of AI in the context of language is that is can tailor the message to the right person at the right time.
One example of a brand doing this is Toyota Mirai, which teamed up with IBM Watson to create thousands of different variations of ad copy based on factors like location, occupation, and past buyer behaviour. The idea was that Toyota wanted to create an ad for every single potential buyer of the car.
In this context, the AI did not replace human creativity. The 50 scripts used to teach the AI were written by humans. However, what it did do was help to create content at scale, as well as use data to create greater relevance and personalisation.
Finally, another area of investment for IBM’s Watson has been programmatic advertising, with the technology being used to further automate the planning, buying, and optimising of advertising.
Using machine learning, Watson learns how campaigns are performing for different audiences, based on different variable such as device and location. It then uses this information to only bid on inventory that aligns with factors that are likely to generate the most success.
It’s no surprise IBM has heavily invested in programmatic. With the help of AI, campaign management not only becomes much more efficient, but far more effective in terms of reaching the right audience at the right time. It’s certainly seen success too. According to reports, IBM has seen a 25% increase in effectiveness in programmatic as a result.
- How machine learning can improve conversion rate & optimise performance
- Big data: the golden prospect of machine learning on business analytics
Don’t forget to download Econsultancy’s ‘Marketer’s Guide to AI and Machine Learning’ report in full