Artificial intelligence (AI) and machine learning are the most exciting developments in business today, they offer many ways to personalise at scale and draw actionable insights from huge pools of customer data.

However, the threats posed to jobs by the advent of AI and machine learning are a serious concern. Any job that primarily entails analysis and processing of data is at risk of being automated out of existence. This is according to Bluegrass Digital managing director Nick Durrant.

AI-powered personalisation seems to offer the prospect of fulfilling – at the level of the individual – the very definition of marketing: creating, communicating, and delivering offerings that have value for customers.

Marketing and merchandising professionals could be forgiven for wondering where this development leaves them. The answer is that AI reaffirms their place in the driver’s seat, fully in control, adding value where AI cannot.

That’s not to deny the likelihood that AI will prompt a significant and undoubtedly large-scale realignment of roles and responsibilities in many organisations. AI drastically reduces the resources required to interpret data, build and manage static segments, deploy and manage internal campaigns or devise content and product permutations.

Conversely there’s a greater role for AI-trained data scientists – whether in-house or third party – who are able to use their technical skills and curiosity to push the boundaries of what AI can do.

Taking a broader view, successful marketing is heavily dependent on context. Applied AI is, by definition, less successful at analysing context. Empathy, creativity and even broader problem-solving skills are missing. The capacity to synthesise business goals, devise strategies, select KPIs and orchestrate tactical marketing efforts: this is all the domain of humans.

Ultimately, the algorithms at the heart of AI-led personalisation are computer programs, written by humans. Determining the right algorithms to use, the desired output, and determining their role in the purchase process is a job that only a marketer or merchandiser can do.

When you combine the expertise of marketers with the algorithms of AI, however, you can present relevant, engaging experiences at a level that was never before possible. Content and campaigns can now be personalised at the individual level. This is important because engagement and conversions are increasingly linked to the quality of your digital experiences.

For commerce companies, great experiences are often what sets them apart from competitors. With AI you can present intelligent campaigns and individualised content that lead to big results.

Companies with a large and growing inventory are particularly well-suited to AI-powered marketing. Near-endless product permutations are possible, far beyond the capabilities of any team.

Whereas merchandisers might previously have matched categories against audiences, for example, a machine learning algorithm will match at the most granular level available: individual against product, informed by every data point held. With the right technology, ingesting new SKUs is immediate, with machine learning driving revenue from day one.

But Average Order Value and other financial metrics aren’t the only KPIs that matter. Marketers and merchandisers bring the insight and knowledge to tailor the presentation of these algorithmically-driven product selections so that other objectives can be met, for example the need to respect commercial agreements around product presentation (don’t show product X against product Y).

The desire to pursue long-term brand equity at the expense of short-term revenue (don’t surface final reductions at the same time as the debut of a new collection). This is achieved through rule engines, weighting and scoring, and prioritisation of campaigns themselves – all of which can be manually configured by those who know their customers, suppliers and market.

The critical role of content marketing throughout the digital experience can be greatly enhanced with AI-led personalisation, addressing questions such as:

Should this user be presented with a brand video or other rich content to drive conversion?

Will an article on accessories in this user’s email reduce the prospect of an abandoned basket?

Do people who search for this particular term want product specs or user generated content?

Content creation is demanding and resource-intensive, but project collaboration and workflow tools can lighten the load. Episerver is a platform that goes one step further by embedding these tools into the CMS itself, reducing the workload still further.

But quality still trumps quantity. Time taken previously to content tag and match is now best used to identify high priority customer journeys and corresponding user segments that stand to gain the most from content - a task made easier through machine learning-led analysis – and create the highest quality content that time and resource affords.

Done well, personalised campaigns can generate a significant return by improving order values, driving down abandoned carts and enriching data that will be used on subsequent visits. Real-time personalisation has much to offer here.

With knowledge of what a visitor has already seen and possibly bought, personalisation powered by AI can serve up banners, recommendations, tailored promotions and rich media to drive conversions. Triggered messages can also be particularly effective.

There are many trigger opportunities available, such when a visitor gives their email in-store, browses certain pages or places specific products in their cart. Messages can be sent via texts, email, with an in-app alert, or via a consumer’s channel of choice.