AI: marketing's friend or foe?

Despite the current hype, artificial intelligence is far from a new phenomenon.

The beginnings of modern AI can be traced to 1956, since when it has powered world champion chess moves, won game shows and been the darling of the science fiction genre.

Interest and investment has gathered significant pace in the last few years leading to new extensions of these early capabilities and integration into everyday life.

At one end of the scale, AI is seen as a key component of solving our big generational challenges with eye-watering incentives for doing so.

On the other, more relatable side, is the likes of Apple’s Siri, Amazon’s Alexa and Google Assistant. Accessible and helpful, the promise of AI’s enormous potential suddenly seems within reach to us all.

Recommendations are more relevant, services are more immediate and better connected, and if an AI skill doesn’t already exist it’s never been easier to create one.

We have certainly taken to our new friends. By 2020 Gartner projects that virtual assistants will power 90% of commercial interactions between people and businesses. As a consequence, the market for artificial intelligence — of which chatbot marketing is a major subset — is expected to become a multi-billion dollar industry over the next decade. VentureBeat published an overview of the market and discovered that there are over 170+ companies, $4 billion in funding, and thousands of bots currently available.

It’s no surprise then that this curiosity about AI is spreading to marketing departments.

Artificial intelligence can already be found undertaking routine marketing tasks to almost perfect standards and with non-exhaustive levels of enthusiasm. Many of us will encounter a marketing virtual agent on a near daily basis when being asked qualifying questions or providing automated answers to standard questions, however their potential goes far beyond this and presents exciting opportunities at every stage of the customer journey. For example:

Big data.

Making sense of data is the cornerstone of Artificial Intelligence marketing. The ‘collect, reason & act’ cycle is put into overdrive with AI adding greater scale, automation and targeting to marketing efforts. People are still incredibly relaxed about granting brands access to their data, allowing it to be collected and integrated from multiple sources to predict what you need before you even realised you needed it. This takes the form of product recommendations, proactive customer service, dynamic pricing and triggered offers.

Assisted search.

AI is quietly reinventing Search. Google’s RankBrain is the deep learning system powering the autosuggest and related searches features that we’ve become accustomed to. It’s also behind the semantic analysis of voice search — determining whether somebody is searching in distress, frustration or curiosity and respond accordingly. Optimising for an AI-first world will become a key requirement for marketing departments and it’ll be brands, not search engines, that customers will blame if the answers to their search are not relevant or up-to-date.

Conversational commerce.

Voice assistants create an opportunity to interject compelling content into everyday situations, such as recipes in the kitchen, linked to an ecommerce platform. It’s this unrivaled integration of Amazon that is keeping marketers up at night.


Banking is one industry embracing AI as a means to getting close to customers again. We stick with our banks out of convenience not loyalty. Banks have become too complacent for their own good with the complexities and barriers to entry providing a lack of incentive to get the basics right. That is until now. Challenger banks are forcing them to up their game and AI can ironically make them appear more human. For example, Swedbank’s Nina new web assistant is engaging in 40,000 customer conversations per month and able to handle more than 350 different customer questions. Of course, banking is far from the only industry in which customer service is lacking.

(AI) will give people the impression that the bank knows them a lot better, and in many ways it will take banking back to the feeling that people had when there were more human interactions.

Alan McIntyre, head of Accenture’s banking practice

Visual commerce.

Thanks to AI and machine learning a customer is now able to take a picture of a jacket they see in the street, match it to a pin on Pinterest and be directed to a retailer who has it in stock. AI is closing the gap between online and offline experience-led commerce.

Media buying.

With online advertising’s integrity issues suddenly out in the open, it’s no wonder that AI media agency Blackwood Seven is making the big networks look over their shoulder. Their AI alternative is giving brands the ability to plan, purchase and optimise their media buying in-house without the usual agency cut.


Creativity is naively seen as immune from AI, however algorithm-driven design and creative direction is already proving to be a cost-effective solution for a lot of businesses. AI-powered logo maker LogoJoy uses machine learning to make it feel like you’re working with a real designer and is already boasting $70,000/month in revenue only one year after launching. While not suitable for all occasions or clients, the AI designer is proving to be a capable, cheap and immediate solution for over 1,000 bootstrapped businesses a month.

Customer service.

Left under human supervision, it’s remarkable how key areas of customer experience have been taken for granted. Recent research has shown that monotonous, repetitive tasks triggers automatic decision making and makes employees more likely to behave unethically.

AI is increasingly stepping in to reduce waiting times, cut costs, close sales, personalise messages and provide the outcomes customers are looking for. Human intervention is still necessary in a lot of those instances — as checks for the algorithms and to make final decisions — meaning our jobs are safe for the time being at least.

Despite all this promise of AI and the fact 80% of marketing executives believe it will revoluntionise marketing by 2020, the reality is that only 10% are currently using AI in their marketing activity.

What’s behind the intention-action gap in marketing departments?

Sure, inertia and lack of technical expertise play a part, but many marketers hold a major question mark over AI’s ability to perform a key part of their role: EMPATHY.

Without empathy it’s impossible to connect with customers on any meaningful level.

Observing, understanding and communicating are arguably the most important traits of any marketer. They feed marketing’s ability to do what we do best: make people feel something. So whilst AI has shown it can be helpful, brands must constantly go beyond this and be thoughtful.

Empathy isn’t just something you feel. It’s what you do with it that counts.

Seeing the world as your customers do, understanding what matters to them, being compelled to act in some way and providing value when you do. It’s what makes a brand lovable. It’s what makes us human.

Perhaps the most famous shortcoming of AI in this area was Target accidentally figuring out a young woman was pregnant by automatically analysing her shopping habits and sending her adverts for baby goods. Problem was that her family hadn’t yet been told the news and found out via Target. Ooops!

Computers can read emotion and can acknowledge it, the problems occur when we need them to react like a machine instead of reacting like a human. This is called ‘affecting computing’ and (for now) it remains a stretch target.

The moment we accept AI devices into our homes and our lives we expect them to display these traits. Research shows that there is a desire for an emotional relationship with AI-equipped devices that goes beyond being an assistant. They become an extension of self.

Simply being fast and correct is no longer enough. Complaints, for example, often need to be dealt with showing sensitivity and with enough authority to sign-off a solution. Conversations with bots will need to flow more freely and build on the last one, giving rise to conversation design as the next big things in UX.

Are our expectations of artificial intelligence too high? This leads to two schools of thought:

Over the last few years, we in the bot community have all been focused on designing user-friendly software, but if we continue to design these bots to just be “friendly” we are missing a million opportunities. We need motivation, empathy, discretion, and relevance.

Toby Barnes, AKQA

The second school thinks AI should stop short of and artificial empathy and be content with being a marketer’s sidekick:

We should resist our predisposition to attribute human traits to our creations and accept these remarkable inventions for what they really are — potent tools that promise a more prosperous and comfortable future.

Jerry Kaplan, Stanford University

There is also the issue of ethics and trust. If AI is trained to be human-like are we deceiving consumers or overstepping accepting levels of intrusion on their personal data? And is it ethical when AI voice is indistinguishable from a human’s?

There’s no easy answer to that questions but it’s clear that the next generation of AI must develop additional qualities as consumers become increasingly comfortable with the technology.

Once AI can master this combination of utility and the human touch marketing will discover it presents a world of opportunities to better serve their customers and find new ones.

More Thoughts