Home Blog


AI-POWERED COPYWRITING: STATE OF THE TECHNOLOGY


April 2021

There has been a lot of excitement about AI-powered copywriting lately, and while very few of our clients have ever broached the topic, we have seen a noticeable hike in the number of online searches for vendors of the technology. Now might thus be a good time to address what may or may not be the elephant in the room for copywriters.

AI and machine learning have already successfully penetrated many sectors of economic activity, from e-commerce to healthcare, marketing, cybersecurity, finance, and logistics. Why not copywriting? Some contend that while demand for copywriting services is growing, supply remains somewhat capped. Copywriters themselves would be delighted if some of their lower-level, repetitive tasks could be automated. After all, sports game summaries, stock reports and even fiction books can already be produced by robots.

Intuitively, one would tend to posit the ability to master a language, think strategically, and create impactful content is so intertwined with what makes us human that delegating such high-level tasks to a computer is unconceivable. As of today, that intuition happens to be mostly correct despite the unfathomable computing power and mass of data at our disposal. Nonetheless, there are areas where AI-powered copywriting already does add tangible value to the communications of both public and private organisations, albeit often of the larger type.

Current AI-powered copywriting tools, which compose human-like copy and determine the version most likely to work out, tend to focus, and perform well, on generating quality short-form content, such as subject lines, headlines, and body of SMS, emails, push notifications and ads. Some have demonstrated their prowess in this area with content that triggers significantly more clicks and engagement – in more than 90% of the time, according to New York-based Persado, whose clients can see for themselves the boost to their campaigns’ performance through its platform.

As London-based Phrasee’s CEO Parry Malm puts it, “AI-powered copywriting has been adopted by creative teams at some of the world’s biggest brands to optimize language at scale, freeing their time to focus on higher concept creative thinking and campaigns.” The words “at scale” are important here and refer to the colossal volume of short-form communications that today permeate every single moment of our lives and represent an ever-growing activity for marketing teams around the world, but also constitutes a gold mine for statistical analysis and machine learning.

For the foreseeable future, long copy such as press releases, articles, newsletters, and magazines, on which AI still performs rather poorly due to the intricacy of logically weaving a string of complex arguments and relevant information into a cohesive piece of writing, remains the preserve of human copywriters. While AI performs quite well on short content due to the massive amount of similar reference materials its algorithms can be trained on, identifying patterns and unearthing value, it will take quite a while before Google is able to produce a 1,500-word insightful and perfectly written article on the various roles a specific company has played to address climate change and why.

Back to our push notifications and mobile ads, the process involved in coming up with turbocharged content is extremely simple. While the steps differ from one provider to the next, for many services like those offered by copy.io, all a user needs to submit is a simple brief that may include a draft headline, body text and CTA, as well as information about the brand (at more established companies like Phrasee, which works in more than 20 languages, the process upon signing up a new client is more comprehensive and involves understanding the brand guidelines, learning about historical campaigns, and sharing questionnaires).

Within a few seconds, a collection of possible copies pops up, ranked by adequacy and/or likely positive impact on human readers, which can then be reviewed by the user. The base copy and one or few of the copies generated can then be fed into the company’s marketing campaign, the results of which, provided it targets a relatively large number of recipients, will inform of the most effective headline, copy, and CTA, a verdict that is usually in favour of the AI.

Besides choosing the seemingly most effective way to craft a short-form copy, an AI is also able to seamlessly and instantly create multiple versions of the same ad or email to target individual recipients according to their personal details and the history of their interactions with the brand. It can also, in another example borrowed from the financial services industry, compose both more tactful and more effective messages that help an organisation successfully and more quickly recover the debts owed to it by customers.

Such capabilities are best harnessed by larger corporations - which are coincidentally the clients of choice of proven technology providers - to derive meaningful benefits. To illustrate this point, Phrasee a few years ago would require several weeks to build a dedicated AI for just one segment and one client. The barriers to entry are indeed quite high, and the most prominent vendors do not provide easy access to a free trial nor propose standard subscription models.

For those more affluent corporations, higher conversion ratios translate into real dollar gains, which can be substantial at scale. AI-powered copywriting was adopted by banking giant JP Morgan Chase, who signed a five-year deal with Persado in 2019, to be applied to its marketing messages, such as emails. In an interview with the Wall Street Journal, the bank reveals how by having the AI make simple, but not necessarily intuitive, changes to an email about equity loans, the number of applications nearly doubled. It was later reported that JP Morgan Chase witnessed a 450% uptick in CTR (click-through rate) in a trial conducted with a different, London-based developer.

A little fact unknown to most of the public, many other industry behemoths, such as Vodafone, Mastercard, Dell, American Express, Virgin Atlantic, Walgreens, and eBay, have already employed the technology for years and thanks to it have consistently and successfully enhanced engagement with customers, smoothly navigated the pandemic, boosted their marketing ROI through better and more consistent language, and often outclassed their competitors.

While the economic benefits for a large organisation far outweigh the cost of deploying the technology, and independent analysts have been quick to write off copywriting as a safe career path, both vendors and their clients do not anticipate the change to cause any reduction in their marketing staff. They instead see AI as enabling their writers to focus on critical and strategic thinking and high-value tasks, emphasising that human touch will always be essential, and a healthy dose of criticism warranted to control the output of an AI.

Looking for quality writing? - Hire Wordcraft, a leading copywriting agency in Hong Kong.


What’s next?

Hopefully this article will have brought you up to date on the latest state of AI-powered copywriting, its benefits, promises, and current limitations. Please let us know if your company is considering trying out the technology and following in the steps of other industry leaders who adopted some of the tools now available to take their communications to the next level.