Once an empty buzzword fuelled by hype and hyperbole, artificial intelligence has now emerged as a marketing necessity, powering countless real-world applications. Its scope is vast and most excitingly, in my view, brings potential to power a critical, much-needed shift from product- to people-driven marketing.
We are close to exhausting the old model of marketing: product-first, optimised across limited dimensions, with linear datasets. In a world where consumers habitually juggle content streams across devices, we can no longer rely on brand-led messages reaching them at predictable moments.
It is time for the industry to embrace people-based marketing, paraphrasing David Williams, Dentsu Aegis Network’s global CEO of CRM and loyalty; it is a model rooted in the knowledge that when customer needs and expectations are met, the result is an affinity that leads to loyalty and even advocacy. People-based marketing succeeds by knowing customers as individual people, not proxies or personas, then marketing to them based on this in-depth knowledge. Instead of dehumanising marketing, AI enables us to do the exact opposite.
AI-driven analytics excel at deciphering convoluted data sets at speed. They enable us to track consumers’ behaviour as they flit across platforms and channels, facilitating a more holistic media planning and buying approach that aligns with media habits.
Where media planning has traditionally been characterised by campaign-driven orientation based on rigid marketing calendars, with broad brush segmentations sweeping up people not even in the market – inefficiency, irrelevance and wastage were inevitable. With AI insights, we need no longer make generic offers for the lowest common denominator.
At iProspect, AI propels our output to deliver individual level segmentations based on relationship to the brand and category intent, delivering messages to the right people at the right moment. It also enables far more robust personalisation of creative that effectively meets needs and desires...
The challenges of AI, the opportunity for people
AI can’t do everything, yet. It excels at dealing with large homogenous datasets, be that Google crunching searches or imagery to deliver a personalised response to your queries, or the Azure platform harnessing Cortana data to provide live translations for text messages. However, marketing intelligence is often made up of a combination of heterogenous datasets, some huge and some small, as no AI rules them all.
This is why at iProspect our teams are trained to bridge the gap between craft and science, harnessing the power of machines and the skills of people. iProspect planners and specialists will bring together the AI-generated insights garnered by Google Analytics, using qualitative studies and SEO best practice to rebuild a site transaction journey that tailors messaging with pinpoint accuracy and enhances conversion massively.
I believe the winners will not be those who create their own AI platforms, but those who foster the best interpretive and implementational skillset to action findings. Whilst at iProspect, we have a specific department – Integrated Intelligence, the role of which is to convert these multi-dimensional datasets into actionable insights for our clients – its skillset cannot remain a specialism. This is why we are committed to training our staff to stay ahead of the next stage of performance marketing. Our rigorous training programme harnesses iProspect University, multiple bursaries and training courses run by all our cloud and AI partners.
The future of AI in marketing – from back office to customer facing
The content AI is able to produce, as well as its direct-to-consumer interactions, will increase massively as algorithms become ever smarter.
Creating content at scale in a world where consumers expect personalised, relevant experiences at all times is a significant pressure on traditional modes of production. AI will become less about repurposing existing assets, and more about creating new ones. The generative adversarial network (GAN) has already become adept at computer renders of people that have never existed, which are almost indistinguishable from actual photos. Whilst this has the potential for abuse, it unlocks so much; not only in the sphere of reactive content production, but also for the potential ambition of content to become so much greater. AI will democratise brilliant branded content.
In the customer service arena, chatbots, text and voice recognition allow brands to respond to customers’ hands-free needs ever faster. Language recognition is reaching unprecedented levels. The teams at Microsoft can transcribe voice data on a par with professional stenographers. This means AI’s speech recognition will make automated customer service a joy rather than a pain. Imagine not having to listen to hold music ever again. Now that’s revolutionary.
Another area ripe for AI’s magic touch is product personalisation. We’re already seeing AI-supported dynamic pricing, where prices of products online are determined according to demand. What about dynamic product creation? Personalisation currently means picking a favourite colour or perhaps having a name engraved on to a gift. If people become more willing to share data with brands, we will see products from makeup to nutrition plans formulated specifically for individuals and their genetic makeup.
Are we optimising Pandora’s box?
When you consider that OpenAI researchers have developed an algorithm that its creators deem too dangerous to release, or that deep fakes use AI to make it seem like someone has made a statement that they would never utter and that Adobe can replicate human speech, one would be foolish not to tread with caution.
Even in the world of customer experience, there are concerning signs of AI having gone awry. As an example, Netflix AI effectively lied to customers about films’ content, showing what people wanted to see rather than what they would actually see. A classic example of an algorithm that did not have the right in-built constraints.
As AI gains a stronger foothold, smarter regulation, training and guidelines on how to create models beckon to ensure that unintended consequences do not occur.
AI is purely additive in my mind. It is a technology that, perhaps ironically, can facilitate more human connections and more empathic marketing strategies, but it requires people do this. And it is when you combine AI with the right people and skillsets that you will get fireworks – a one plus one equals three scenario.
AI is here; the onus is now on marketers to plug into its potential.