Augment not replace: why AI is not good enough to become a human replacement

null

Talent and expertise appear to be no guarantee of job security in the twenty-first century. With the rise of Artificial Intelligence (AI), knowledge workers across the board are increasingly at risk of being replaced - from customer services staff being superseded by simple chatbots to algorithms that promise faster, more accurate diagnosis than doctors. Does that include marketing teams? Are content developers and merchandisers also at risk of being sidelined by one of the broad range of technologies now bundled into the AI umbrella?

Zalando’s announcement earlier in the year that it was replacing 250 marketing communications jobs with AI certainly provoked interest but the machine learning algorithms posing as ‘AI’ within marketing today are (for the vast majority) in no way good enough to replace human interaction. They do, however, offer a raft of benefits that could and should transform human activity – releasing individuals from mundane, tedious manual tasks and enabling teams to leverage near real-time insight to prioritise activity and become far more effective and efficient, Duncan Keene, UK Managing Director ContentSquare, insists.

Misguided AI Thinking

AI is incredibly powerful, and it clearly has the capacity to transform virtually every aspect of human activity, over time. But organisations’ persistent focus on the use of AI to replace human beings is flawed; while there will without a doubt be occasions where an AI operating in the background will remove the need for human activity, such opportunities are limited – and potentially limiting. The real focus for organisations should be around utilising AI to accelerate change and transform the speed of response.

Indeed, there is growing recognition that even simple AI such as chatbots and Natural Language tools should not be deployed simply with a view to eradicating human activity but to release humans from mundane, often mind-numbing tasks – such as updating change of address – and enabling them to focus on added value customer interactions.

Why, then, would a high-profile retailer such as Zalando, announce that AI was replacing 250 of its marketing team? What message is that sending to the rest of the business? Or it's customer base?

Aid not Replace

AI is just technology; clever technology but still technology. And over the past few decades organisations have learnt – often painfully – that technology alone is never the answer. It is about culture, people and leveraging the right technologies to address specific business problems. The technology that aids business based AI and automation absolutely has a place in all marketing departments, but it should be used to empower staff to make better, more insightful decisions – not replace them. The downsides of removing intelligent, experienced marketing staff from the mix are significant.

For example, some companies are considering the use of AI to automatically merchandise product list pages.  With so much customer data available, the argument goes, the use of algorithms to automate this process would remove the need for the vast, expensive team of visual merchandisers currently manually co-ordinating pages. But these data sets are limited to customer views and conversions – today’s AI cannot consider cultural trends or include a brand’s preferences for specific product promotion or image. A retailer opting to follow the machine learning route with none of the intelligence provided by human awareness or context will miss out on essential opportunities to build brand identity and reinforce the customer message.

Where, for example, is the promotion of products worn on last night’s Love Island? Or immediate tie-in with weather forecasts, news events or celebrity endorsement? Today’s algorithms are one dimensional; they lack the ability to factor in these essential insights – and that is where the human input is essential.

Science Fiction

And let’s be frank – such intelligence is a world away from the reality of many eCommerce content teams today. Most spend more time reporting than optimising; there is minimal opportunity to dive into the analytics and identify valid opportunities to improve the site. For many, the idea of being to make relevant changes in near real-time is more science fiction than AI itself.

Releasing these individuals from the tedium of reporting and delivering the insight required to prioritise activity, highlight problems and identify valuable changes can transform performance. The focus should not be on replacing individuals, but on leveraging technology such as behavioural analytics to surface job specific insights – such as the performance of the checkout pages on the mobile website for a specific segment of customers. Measuring how many times a frustrated visitor clicks on key Call to Actions (CTA) in the checkout funnel can provide an immediate alert regarding an issue with a critical aspect of the checkout journey. So rather than losing customers who are ready but struggling to complete a purchase, the technology can quickly surface the insight and enable the team to reprioritise and investigate this critical issue immediately.

Similarly, brands are constantly changing website content in an attempt to engage and convert the customer – yet have minimal insight into the value of such changes. Rather than embark upon multiple A/B tests to compare performance, behavioural analytics can provide a rapid understanding of the attractiveness, effectiveness and value of the content.  No more blindly producing content with fingers crossed; with rapid and trusted insight, teams can concentrate on producing content that is proven to be both attractive and effective.

Empower not Replace

Announcing an ‘AI Strategy’ is the wrong way to approach these technologies. Companies need to consider specific problems, assess the skills available and then determine how technologies – including AI – could be used to support, not replace, those individuals to address the problem. Furthermore, the concept of an ‘AI strategy’ sends the wrong message and can undermine employee morale by suggesting job cuts.

AI can of course surface insights far faster than any human; but right now, it is the human who needs to add context to ensure the insights are appropriately actioned. With the right culture and business-led approach, AI can make life so much easier for marketers; eradicating the mundane reporting and delivering valid, an immediate insight that can be used in tandem with their expertise and experience, their understanding of current events and trends, to support ever better and faster decision making.  

Essentially, AI should augment, not replace, human experience. 

Duncan Keene, UK Managing Director of ContentSquare 

Image Credit: Geralt / Pixabay